Digital rights management engine systems and methods

[0001] This application claims the benefit of U.S. Provisional Application No. 60/728,089, filed Oct. 18, 2005, U.S. Provisional Application No. 60/772,024, filed Feb. 9, 2006, U.S. Provisional Application No. 60/744,574, filed Apr. 10, 2006, U.S. Provisional Application No. 60/791,179, filed Apr. 10, 2006, U.S. Provisional Application No. 60/746,712, filed May 8, 2006, U.S. Provisional Application No. 60/798,925, filed May 8, 2006, and U.S. Provisional Application No. 60/835,061, filed Aug. 1, 2006. U.S. Provisional Application Nos. 60/728,089, 60/772,024, 60/744,574, 60/791,179, 60/746,712, 60/798,925, and 60/835,061 are incorporated herein by reference in their entirety for any purpose.

[0002] A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

[0003] In modern computing systems, it is often desirable to limit access to electronic content, services, and/or processing resources, and/or to allow only certain entities to perform certain actions. A variety of techniques have been developed or proposed to enable such control. These techniques are often referred to as digital rights management (DRM) techniques because, in general terms, their goal is to manage the rights of various entities in digital or other electronic content, services, or resources. A problem with many prior art techniques is that they are overly complex, overly restrictive, relatively inflexible, fail to enable certain natural types of relationships and processes, and/or are uninteroperable with other DRM systems.

[0004] Systems and methods are described herein that can be used to ameliorate some or all of these problems. It should be appreciated that embodiments of the presently described inventive body of work can be implemented in numerous ways, including as processes, apparatuses, systems, devices, methods, computer readable media, and/or as a combination thereof. Several illustrative embodiments are described below.

[0005] The inventive body of work will be readily understood by referring to the following detailed description in conjunction with the accompanying drawings, in which:

[0006] FIG. 1 shows an illustrative system for managing the use of electronic content.

[0007] FIG. 2 shows a more detailed example of a system that could be used to practice embodiments of the inventive body of work.

[0008] FIG. 3 shows how an illustrative digital rights management (DRM) engine might function in a network that uses DRM.

[0009] FIG. 4 shows a collection of nodes and links used to model the relationships in a DRM system.

[0010] FIG. 5 is a flowchart illustrating how an embodiment of a DRM engine might determine whether a requested action is authorized.

[0011] FIG. 6 shows an example of a DRM license in accordance with one embodiment of the inventive body of work.

[0012] FIGS. 7A and 7B illustrate the use of agents in one embodiment.

[0013] FIG. 8 shows an example of a DRM license.

[0014] FIG. 9 is a more detailed example of how a DRM engine might determine whether a requested action is authorized.

[0015] FIG. 10 is a more detailed example of how a DRM engine executes a control program in one embodiment object.

[0016] FIG. 11 shows an illustrative embodiment DRM engine running on a device.

[0017] FIG. 12 is a flowchart illustrating the steps involved in executing a control program in one embodiment.

[0018] FIG. 13 shows the elements that make up a content consuming client application in one embodiment.

[0019] FIG. 14 shows the elements that make up a content packaging application in one embodiment.

[0020] FIG. 15 shows a key derivation mechanism in accordance with one embodiment.

[0021] FIG. 16 shows an example of a DRM system.

[0022] FIG. 17 shows an example of a DRM system that provides for temporary login.

[0023] FIG. 18 shows the high-level architecture of an illustrative system for managing enterprise documents.

[0024] FIG. 19 shows an example of a how a system such as that shown in FIG. 18 can be used to manage access to or other use of a document.

[0025] FIG. 20 shows an additional example of a how a system such as that shown in FIG. 18 can be used to manage access to or other use of a document.

[0026] FIG. 21 shows additional features of the example shown in FIG. 20.

[0027] FIG. 22 shows another illustrative system for managing electronic content within an enterprise.

[0028] FIG. 23 illustrates how the systems and methods described herein could be applied to manage healthcare records.

[0029] FIG. 24 is an illustration of how the systems and methods presented herein could be used in a context of an electronic subscription service.

[0030] FIG. 25 is an illustration of how the systems and methods described herein could be used in a context of a home network domain.

[0031] FIG. 26 illustrates the interactions that take place between a host application and a DRM client engine in one example embodiment.

[0032] FIG. 27 illustrates the interactions that take place between a host application and a packaging engine in one illustrative embodiment.

[0033] FIG. 28A is a more detailed illustration of a license in accordance with one embodiment.

[0034] FIG. 28B illustrates the relationship between links and nodes in one example embodiment.

[0035] FIG. 29 illustrates the operating environment of an illustrative implementation of a virtual machine.

[0036] FIG. 30 illustrates an extended status block data structure in accordance with one embodiment.

[0037] FIG. 31A shows a memory image of a data segment in one embodiment.

[0038] FIG. 31B shows an example of the memory image of a code segment in one embodiment.

[0039] FIG. 31C shows an example of an export entry memory image in one embodiment.

[0040] FIG. 31D shows a generic example of an export table entry in one embodiment.

[0041] FIG. 31E shows an example of an export table entry for an example entry point.

[0042] FIG. 32 shows an example of a license transfer protocol.

[0043] FIG. 33 shows another example of a license transfer protocol in accordance with one embodiment.

[0044] FIG. 34 shows a mechanism for protecting the integrity of license objects in one embodiment.

[0045] FIG. 35 shows a mechanism for protecting the integrity of license objects in another embodiment.

[0046] FIG. 36 illustrates a proximity checking protocol in accordance with one embodiment.

[0047] FIG. 37 illustrates the use of a proximity check protocol in accordance with one embodiment.

[0048] FIG. 38 illustrates an interaction between a client and a license server in one embodiment.

[0049] FIG. 39 is more detailed illustration of an interaction between a client and a license server in one embodiment.

[0050] FIG. 40 shows an example of an entity with multiple roles.

[0051] FIG. 41 illustrates a bootstrap protocol in accordance with one embodiment.

[0052] FIG. 42 shows the relationship between c14n-ex and an illustrative XML canonicalization in one embodiment.

[0053] A detailed description of the inventive body of work is provided below. While several embodiments are described, it should be understood that the inventive body of work is not limited to any one embodiment, but instead encompasses numerous alternatives, modifications, and equivalents. In addition, while numerous specific details are set forth in the following description in order to provide a thorough understanding of the inventive body of work, some embodiments can be practiced without some or all of these details. Moreover, for the purpose of clarity, certain technical material that is known in the related art has not been described in detail in order to avoid unnecessarily obscuring the inventive body work.

[0054] Commonly-assigned U.S. patent application Ser. No. 10/863,551, Pub. No. 2005/0027871 A1 (“the '551 application”), which is hereby incorporated by reference, describes embodiments of a digital rights management (DRM) architecture and a novel DRM engine that overcome some of the weaknesses that characterize many previous DRM implementations. The present application describes enhancements, extensions, and modifications to, as well as alternative embodiments of, the architecture and DRM engine described in the '551 application, as well as new components, architectures, and embodiments. It will thus be appreciated that the material described herein can be used in the context of an architecture and/or DRM engine such as that described in the '551 application, as well as in other contexts.

1. Example DRM System

[0055] FIG. 1 shows an illustrative system 100 for managing electronic content. As shown in FIG. 1, an entity 102 holding rights in electronic content 103, packages the content for distribution and consumption by end users 108a-e (referred to collectively as “end users 108,” where reference numeral 108 refers interchangeably to the end user or the end user's computing system, as will be clear from the context). For example, entity 102 may comprise a content owner, creator, or provider, such as a musician, movie studio, publishing house, software company, author, mobile service provider, Internet content download or subscription service, cable or satellite television provider, the employee of a corporation, or the like, or an entity acting on behalf thereof, and content 103 may comprise any electronic content, such as digital video, audio, or textual content, a movie, a song, a video game, a piece of software, an email message, a text message, a word processing document, a report, or any other entertainment, enterprise, or other content.

[0056] In the example shown in FIG. 1, entity 102 uses a packaging engine 109 to associate a license 106 with the packaged content 104. License 106 is based on the policies 105 or other wishes of entity 102, and specifies permitted and/or prohibited uses of the content and/or one or more conditions that must be satisfied in order to make use of the content, or that must be satisfied as a condition or consequence of use. The content may also be secured by one or more cryptographic mechanisms such as encryption or digital signature techniques, for which a trust authority 110 may be used to obtain the appropriate cryptographic keys, certificates, and/or the like.

[0057] As shown in FIG. 1, packaged content 104 and licenses 106 can be provided to end users 108 by any suitable means, such as via a network 112 like the Internet, a local area network 103, a wireless network, a virtual private network 107, a wide area network, and/or the like, via cable, satellite, broadcast, or cellular communication 114, and/or via recordable media 116 such as a compact disc (CD), digital versatile disk (DVD), a flash memory card (e.g., an Secure Digital (SD) card), and/or the like. Packaged content 104 can be delivered to the user together with license 106 in a single package or transmission 113, or in separate packages or transmissions received from the same or different sources.

[0058] The end user's system (e.g., a personal computer 108e, a mobile telephone 108a, a television and/or television set-top box 108c, a portable audio and/or video player, an eBook reader, and/or the like) contains application software 116, hardware, and/or special-purpose logic that is operable to retrieve and render the content. The user's system also includes software and/or hardware, referred to herein as a digital rights management engine 118, for evaluating the license 106 associated with the packaged content 104 and enforcing the terms thereof (and/or enabling application 116 to enforce such terms), such as by selectively granting the user access to the content only if permitted by the license 106. Digital rights management engine 118 may be structurally or functionally integrated with application 116, or may comprise a separate piece of software and/or hardware. Alternatively, or in addition, a user's system, such as system 108c, may communicate with a remote system, such as system 108b, (e.g., a server, another device in the user's network of devices, such as a personal computer or television set-top box, and/or the like) that uses a digital rights management engine to make a determination 120 as to whether to grant the user access to content previously obtained or requested by the user.

[0059] The digital rights management engine, and/or other software on the user's system, or in remote communication therewith, may also record information regarding the user's access to or other use of the protected content. In some embodiments, some or all of this information might be communicated to a remote party (e.g., a clearinghouse 122, the content creator, owner, or provider 102, the user's manager, an entity acting on behalf thereof, and/or the like), e.g., for use in allocating revenue (such as royalties, advertisement-based revenue, etc.), determining user preferences, enforcing system policies (e.g., monitoring how and when confidential information is used), and/or the like. It will be appreciated that while FIG. 1 shows an illustrative DRM architecture and a set of illustrative relationships, the systems and methods described herein can be practiced in any suitable context, and thus it will be appreciated that FIG. 1 is provided for purposes of illustration and explanation, not for purposes of limitation.

[0060] FIG. 2 shows a more detailed example of a system 200 that could be used to practice embodiments of the inventive body of work. For example, system 200 might comprise an embodiment of an end user's device 108, a content provider's device 109, and/or the like. For example, system 200 may comprise a general-purpose computing device such as a personal computer 108e or network server 105, or a specialized computing device such as a cellular telephone 108a, personal digital assistant, portable audio or video player, television set-top box, kiosk, gaming system, or the like. System 200 will typically include a processor 202, memory 204, a user interface 206, a port 207 for accepting removable memory 208, a network interface 210, and one or more buses 212 for connecting the aforementioned elements. The operation of system 200 will typically be controlled by processor 202 operating under the guidance of programs stored in memory 204. Memory 204 will generally include both high-speed random-access memory (RAM) and non-volatile memory such as a magnetic disk and/or flash EEPROM. Some portions of memory 204 may be restricted, such that they cannot be read from or written to by other components of the system 200. Port 207 may comprise a disk drive or memory slot for accepting computer-readable media 208 such as floppy diskettes, CD-ROMs, DVDs, memory cards, SD cards, other magnetic or optical media, and/or the like. Network interface 210 is typically operable to provide a connection between system 200 and other computing devices (and/or networks of computing devices) via a network 220 such as the Internet or an intranet (e.g., a LAN, WAN, VPN, etc.), and may employ one or more communications technologies to physically make such connection (e.g., wireless, Ethernet, and/or the like). In some embodiments, system 200 might also include a processing unit 203 that is protected from tampering by a user of system 200 or other entities. Such a secure processing unit can help enhance the security of sensitive operations such as key management, signature verification, and other aspects of the digital rights management process.

[0061] As shown in FIG. 2, memory 204 of computing device 200 may include a variety of programs or modules for controlling the operation of computing device 200. For example, memory 204 will typically include an operating system 220 for managing the execution of applications, peripherals, and the like; a host application 230 for rendering protected electronic content; and a DRM engine 232 for implementing some or all of the rights management functionality described herein. As described elsewhere herein, DRM engine 232 may comprise, interoperate with, and/or control a variety of other modules, such as a virtual machine 222 for executing control programs, and a state database 224 for storing state information for use by virtual machine 222, and/or one or more cryptographic modules 226 for performing cryptographic operations such as encrypting and/or decrypting content, computing hash functions and message authentication codes, evaluating digital signatures, and/or the like. Memory 204 will also typically include protected content 228 and associated licenses 229, as well as cryptographic keys, certificates, and the like (not shown).

[0062] One of ordinary skill in the art will appreciate that the systems and methods described herein can be practiced with computing devices similar or identical to that illustrated in FIG. 2, or with virtually any other suitable computing device, including computing devices that do not possess some of the components shown in FIG. 2 and/or computing devices that possess other components that are not shown. Thus it should be appreciated that FIG. 2 is provided for purposes of illustration and not limitation.

[0063] A digital rights management engine and related systems and methods are described herein that can be used to provide some or all of the rights management functionality of systems such as those shown in FIGS. 1 and 2, or in other types of systems. In addition, a variety of other systems and methods are described below that could be used in the context of systems such as those shown in FIGS. 1 and 2, as well as in other contexts, including contexts unrelated to digital rights management.

2. DRM Engine Architecture

[0064] In one embodiment a relatively simple, open, and flexible digital rights management (DRM) engine is used to implement core DRM functions. In a preferred embodiment, this DRM engine is designed to integrate relatively easily into a web services environment such as that described in the '551 application, and into virtually any host environment or software architecture. In a preferred embodiment, the DRM engine is independent of particular media formats and cryptographic protocols, allowing designers the flexibility to use standardized or proprietary technologies as required by the particular situation. The governance model used by preferred embodiments of the DRM engine is simple, but can be used to express sophisticated relationships and business models.

[0065] Some of the illustrative embodiments of a DRM engine that are described below relate to an example implementation referred to as “Octopus”; however, it will be appreciated that the present inventions are not limited to the specific details of the Octopus example, which are provided for purposes of illustration, not limitation.

[0066] 1.1. Overview

[0067] FIG. 3 shows how an illustrative DRM engine 303a might function in a system 302 that uses DRM. As shown in FIG. 3, in one embodiment DRM engine 303a is embedded or integrated within a host application 304a (e.g., a content rendering application such as an audio and/or video player, a text-rendering application such as an email program, word processor, eBook reader, or document reader, and/or the like) or is in communication therewith. In one embodiment, DRM engine 303a performs DRM functions and relies on host application 304a for services such as encryption, decryption, file management, and/or other functions can be more effectively provided by the host. For example, in a preferred embodiment DRM engine 303a is operable to manipulate the DRM objects 305 which comprise a license 306 to protected content 308. In some embodiments, DRM engine 303a may also delivers keys to host application 304a. As shown in FIG. 3, either or both of DRM engine 303a and host application 304a may make use of web services 305a and/or host services 306a for processing and/or information needed to complete their respective tasks. The '551 application provides examples of such services, and the manner in which a DRM engine 303a and host application 304a might interoperate therewith.

[0068] In the example shown in FIG. 3, DRM engine 303a, host application 304a, host services 306a, and web services interface 305a are loaded onto a device 300a, such as an end user's personal computer (PC). Device 300a is communicatively coupled to a server 300b, from which content 308 and license 306 were obtained, as well as a portable device 300d, to which device 300a may forward content 308 and/or license 306. Each of these other devices may include a DRM engine 303 that is similar or identical to DRM engine 300a, which can be integrated with the particular host application and host environment of the device. For example, server 300b might include a host application 304b that performs bulk packaging of content and/or licenses, and makes use of a DRM engine 300a to evaluate controls associated with the content that is being packaged in order to comply with any redistribution restrictions. Similarly, device 300c might include a host application 304c that is capable of both rendering and packaging content, while device 300a might include a host application that is simply able to render content. As yet another example of the potential diversity of host environments, device 300d might not include a web services interface, but may instead rely on communication with device 300a, and web services interface 305a to the extent host application 304d and/or DRM engine 303d require the use of any web services. FIG. 3 is only one example of a system in which a DRM engine might be used; it will be appreciated that embodiments of the DRM engines described herein can be implemented and integrated with applications and systems in many different ways, and are not limited to the illustrative examples shown in FIG. 3.

[0069] 1.2. Objects

[0070] In preferred embodiments, content protection and governance objects are used to represent entities in a system, to protect content, to associate usage rules with the content, and to determine if access can be granted when requested.

[0071] As described in more detail below, in one embodiment, the following objects are used:

	Object Type	Function
	Node	Represents entities
	Link	Represents a directed relationship between
		entities
	Content	Represents content (e.g., media content)
	ContentKey	Represents encryption keys used to encrypt
		content
	Control	Represents usage rules that govern interaction
		with content
	Controller	Represents associations between Control and
		ContentKey objects
	Protector	Represents associations between Content and
		ContentKey objects

[0072] 1.2.1. Node Objects

[0073] Node objects are used to represent entities in the system. In practice, a node will usually represent a user, a device, or a group. Node objects will also typically have associated attributes that represent certain properties of the entity associated with the node.

[0074] For example, FIG. 4 shows two users (Xan 400 and Knox 402), two devices (PC 404 and portable device 406), and several entities that represent groups (e.g., members of the Carey family 408, members of the public library 410, subscribers to a particular music service 412, RIAA-approved devices 414, and devices manufactured by a specific company 416), each having an associated node object.

[0075] In one embodiment node objects include attributes that define what the node represents. One example of an attribute is a node type. Besides representing users, groups, or devices, the node type attribute could be used to represent other entities. In some embodiments, a node object can also include cryptographic key information, such as when an embodiment of the key derivation and distribution techniques described elsewhere herein is used.

[0076] In some embodiments, node objects also include a confidentiality asymmetric key pair that is used for targeting confidential information to the subsystems that have access to the confidential parts of the node object. This could be the entity that the node represents (for example, the Music Service 412) or some entity responsible for managing the node (for example, the end user (e.g., Knox 402) could be responsible for managing his or her portable device 406).

[0077] 1.2.2. Link Objects

[0078] In a preferred embodiment, link objects are signed objects used to show the relationship between two nodes. For example, in FIG. 4 the link 418 from the PC node 404 to Knox 402 shows ownership. The link from Knox 402 to the Carey family node 408 shows membership, as does the link from the Carey family node 408 to the Music Service Subscribers node 412. In one embodiment, link objects express the relationship between two nodes, and thus the relationships shown in FIG. 4 could be represented using ten links.

[0079] As shown in FIG. 4, a graph 420 can be used to express the relationship between nodes, where link objects are the directed edges between nodes. For example, in FIG. 4, the relationship between the Carey family node 408 and the Music Service node 412 asserts that there exists a directed edge 422 in the graph whose vertices are the Carey family node 408 and the Music Service node 412. Knox 402 and Xan 400 are members of the Carey family 408. Because Knox 402 is linked to the Carey family 408 and the Carey family 408 is linked to the Music Service 412 there is said to be a path between Knox 402 and the Music Service 412. A DRM engine considers a node to be reachable from another node when there is a path from that node to the other node. This allows a control to be written that allows permission to access protected content based on the condition that a node is reachable from the device where the application that requests access to the protected content is executing.

[0080] As described in more detail below, link objects can also optionally contain some cryptographic data that allows derivation of content keys. Link objects may also contain control programs that define the conditions under which the link may be deemed to be valid. Such control programs can be executed or interpreted (these terms are used interchangeably herein) by a DRM engine's virtual machine to evaluate the validity of a link (e.g., to determine whether the link may be used to reach a given node in an authorization graph).

[0081] In one embodiment, links are signed. Any suitable digital signature mechanism can be used, and in one embodiment the DRM engine does not define how the link objects are signed and does not evaluate any associated certificates, instead, it relies on the host system to verify any such signatures and/or certificates. This allows the system architect or administrator to define the lifetime of a link object, to revoke it, and so on (e.g., by using expiring keys or certificates, revocation, and/or the like), thus providing an additional layer of policy management and security on top of the policy management and security provided by the DRM engine's evaluation of control programs and DRM objects in the context of specific pieces of protected content and/or links (for example, expiration of a link could alternatively, or in addition, be implemented by including an appropriate control program in the link object itself, which, when executed would enforce the expiration date or other validity period). In one embodiment, the DRM engine is generic, and works with any suitable encryption, digital signature, revocation, and/or other security scheme that is used by the host application and/or environment. Thus, for example, if the DRM engine needs to determine if a particular link has been properly signed, it might simply call the host application (and/or a host or system cryptographic service) to verify the signature in accordance with the particular signature scheme chosen by the system designer, the details of which the DRM engine itself may be unaware. In other embodiments, the DRM engine itself performs the actual signature evaluation, relying on the host simply to indicate the appropriate signature algorithm to use.

[0082] 1.2.3. Content Protection and Governance

[0083] Referring once again to FIG. 3, in a typical scenario, a content provider 300b uses an application 304b that includes a packaging engine to encrypt or otherwise cryptographically secure a piece of electronic content 308 and creates a license 306 that governs access to or other use of that content. In one embodiment, license 308 comprises a set of objects 305 that specify how content 308 may be used, and also includes the content's encryption key(s) and/or the information needed to obtain them. In one embodiment, content 308 and license 306 are logically separate, but are bound together by internal references (e.g., using object IDs 310). In many situations it may be convenient to store and/or deliver the content and the license together; however, this is not required in preferred embodiments. In one embodiment, a license can apply to more than one item of content, and more than one license can apply to any single item of content.

[0084] As shown in FIG. 3, when a host application 304a running on a client device 300a wants to perform an action on a particular piece of content 308, it asks DRM engine 303a to check if the action it intends to perform (e.g., “play”) is allowed. In one embodiment, the DRM engine 303a will, from the information contained in the objects 305 comprising content license 306, load and execute a control program associated with content 308, and permission to perform the action will be granted or denied based on the result returned by the control program. Permission will typically require that some conditions be met, such as the condition that a node be reachable from the node representing the requesting entity/device 300a.

[0085] FIG. 5 is a flowchart illustrating how an embodiment of a DRM engine might determine whether a requested action (e.g., viewing a piece of content) is authorized. As shown in FIG. 5, a request to evaluate a license for a given action is received (500). For example, this request might be received from the host application, after the host received a request from a user to perform the specified action. As shown in FIG. 5, the DRM engine evaluates the specified license (502), and determines whether the requested action is authorized (504). For example, the license may contain a control program that the DRM engine executes, the output of which is used to make the authorization decision. If the license authorizes the requested action (i.e., a “yes” exit from block 504), then the DRM engine indicates to the host application that the request is granted (506). Otherwise, the DRM engine indicates to the host application that the request is denied (508). In some embodiments, the DRM engine may also return to the host application a variety of metadata that e.g., associates conditions with a grant of authorization (e.g., obligations and/or callbacks), or provides additional information regarding the cause of a denial of authorization. For example, the DRM engine may indicate that the requested action is allowed only if the host application logs certain information regarding performance of the requested action, or as long as the host application calls the DRM engine back at predefined time intervals to, e.g., re-evaluate the license. Additional information on such obligations, callbacks, and other metadata returned by the DRM engine is provided below. If the requested action is authorized, the content key will be retrieved (e.g., from the license's ContentKey object), and used to release the content for the requested use.

[0086] 1.2.4. License DRM Objects

[0087] As shown in FIG. 6, in preferred embodiment a license 600 is a collection of objects. In the example shown in FIG. 6, license 600 comprises a ContentKey object 602, a protector object 604, a controller object 606, and a control object 608. As shown in FIG. 6, ContentKey object 602 includes encrypted key data 610 (e.g., an encrypted version of the key needed to decrypt encrypted content item 612) and information regarding the cryptosystem used to encrypt the key data. Protector object 604 binds ContentKey object 602 to one or more content objects 614. As shown in FIG. 6, control object 608 includes and protects a control program 616 that specifies how content object 614 is governed. In a preferred embodiment, control program 616 is a piece of executable bytecode that runs on a virtual machine operated by the DRM engine. The control program governs whether certain actions can be performed on the content by checking for satisfaction of conditions specified in the control program, such as whether certain nodes are reachable using valid link objects, whether certain state objects have been stored, whether the host environment has certain characteristics, and/or the like. Referring once again to FIG. 6, controller object 606 is used to bind one or more ContentKey object 602 to control object 608.

[0088] License 600 may also comprise additional objects, such as metadata providing a machine- or human-readable description of the content-access conditions required by the license. Alternatively, or in addition, such metadata can be included as a resource extension of one of the other objects (e.g., control object 608). In the embodiment shown in FIG. 6, control object 608 and controller object 606 are both signed, so that the system can verify that the control information is from a trusted source before using it to make content-access decisions. In one embodiment, the validity of control object 608 can also be checked through verification of a secure hash included in controller object 606. Controller object 606 can also contain a hash value for each of the keys or other key data contained in the ContentKey object(s) 602 that it references, thereby rendering it relatively difficult for an attacker to tamper with the binding between the key data and the ContentKey object.

[0089] As shown in FIG. 6, in one embodiment content 612 is encrypted and is included in a content object 614. The decryption key 610 that is used is included within (or referenced by) ContentKey object 602, and the binding between the two is represented by the protector object 604. As shown in FIG. 6, unique IDs are used to facilitate the binding between content object 614 and ContentKey object 602. The rules that govern the use of key 610 to decrypt content 612 are included within control object 608, and the binding between control object 608 and ContentKey 602 is represented by controller object 606, again using unique IDs.

[0090] It will be appreciated that while FIG. 6 shows the objects that comprise a license in one preferred embodiment, the DRM systems and methods described herein are not limited to the use of this license structure. For example, without limitation, licenses could be used in which the functionality of the various objects shown in FIG. 6 are combined in a smaller number of objects, or spread out over additional objects, or broken up between objects in a different manner. Alternatively, or in addition, embodiments of the systems and methods described herein can be practiced with licenses that lack some of the functionality enabled by the license structure shown in FIG. 6, and/or that provide additional functionality. Thus it will be appreciated that any suitable mechanism for associating licenses with content can be used in accordance with the principles described herein, although in preferred embodiments the advantageous structure shown in FIG. 6 is used.

[0091] 1.3. State Database

[0092] In one embodiment, the DRM engine includes, or has access to, a secure, persistent object store that can be used to provide a secure state storage mechanism. Such a facility is useful to enable control programs to be able to read and write state information that is persistent from invocation to invocation. Such a state database can be used to store state objects such as play-counts, date of first use, accumulated rendering times, and/or the like, as well as membership status, and/or any other suitable data. In some embodiments, a DRM engine executing on a first system may not have access to a local state database, and may be operable to access a remote state database, e.g., using web and/or host services. In some situations, it may be necessary for a DRM engine executing on a first system to access state information stored in a database on a remote system. For example the first system may not include a state database, or may not have the information it needs in its own state database. In some embodiments, when a DRM engine is faced with such a situation, it might access a remote state database via a services interface, and/or by using agent programs, as described in more detail below.

[0093] 1.4. About Control Programs

[0094] The systems and methods described herein make use of control programs in a variety of contexts. For example, control programs contained in control objects can be used to express the rules and conditions governing the use of protect content. In addition, control programs in link objects can be used to express the rules and conditions used to determine whether the link is valid for a given purpose (e.g., a node reachability analysis). Such control programs are sometimes referred to herein as link constraints. Yet another context in which control programs may be used is in agent or delegate objects, were the control code is used to perform an action on behalf of another entity (in the case of agent control programs) or on behalf of another control (in the case of delegate control programs).

[0095] In one embodiment, control programs are executed or interpreted by a virtual machine hosted by a DRM engine, as opposed to being executed directly by a physical processor. It will be appreciated, however, that a physical processor or other hardware logic could be readily constructed to execute control programs. In one embodiment, the control programs are in byte-code format, which facilitates interoperability across platforms.

[0096] In a preferred embodiment, control programs are written in assembly language and converted into byte code by an assembler program. In other embodiments, templates and/or high-level rights expression languages could be used to provide the initial expression of rights, rules, and/or conditions, and a compiler could be used to convert the high-level expression into byte code for execution by an embodiment of the DRM engine described herein. For example, rights expressions written in a proprietary DRM format could, with an appropriate compiler, be converted or translated into a functionally equivalent byte code expression for execution on an embodiment of the DRM engine described herein, thus enabling a protected piece of content to be used, in accordance with the conditions specified by the content provider, on systems that understand the proprietary DRM format, as well as systems that included a DRM engine such as that described herein. It should also be appreciated that the digital rights management engine systems and methods described herein are not limited to the use of byte code rights expressions, interpreted by a virtual machine. Instead, in some embodiments, rights can be expressed in any suitable manner (e.g., using a high-level rights expression language (REL), a template, etc.), and the authorization graph and/or other techniques described herein performed using an application program designed to recognize and evaluate such rights expressions.

[0097] 1.4.1. Conditions

[0098] As previously indicated, control programs typically express one or more conditions that must be satisfied in order for a request to use a piece of content to be granted, for a link to be deemed valid, and/or the like. Any suitable conditions can be used, depending on the requirements of the content provider or system architect, and/or the functionality provided by the system.

[0099] In preferred embodiments, the virtual machine used by the DRM engine supports arbitrarily complex programs that are capable of testing for conditions such as some or all of the following:

• • [0100] Time-based conditions: Comparing a client time value to a value or values specified in the control program.
  • [0101] Targeting a particular node: Checking whether a certain node is reachable from another node. This concept provides support for such models as domains, subscriptions, memberships, and the like.
  • [0102] Testing if certain node attributes match specified values: Checking any of a node's attributes, such as, for example, whether the rendering capabilities of a device associated with a node meet fidelity requirements.
  • [0103] Testing if the security-related metadata at the client is up-to-date: Checking, for example, whether the client has an acceptable version of the client software and an accurate measure of time. In some embodiment, such a check might rely, for example, on assertions in one or more certificates from a data certification service.
  • [0104] State-based conditions: Checking information in the state database. For example, the state database may contain information generated as a result of previous execution of control programs, and/or tokens attesting to ownership of subscriptions, membership, and/or the like, thereby enabling evaluation of conditions involving counters (e.g., number of plays, number of exports, elapsed time limits, etc.) and other information regarding recorded events and conditions.
  • [0105] Environmental characteristics: For example, checking whether hardware and/or software in the host environment has certain characteristics, such as the ability to recognize and enforce obligations; checking for the presence or absence of certain software or hardware components, such as a secure output channel; checking proximity information, such as the proximity of a requesting device to another device or application; checking the characteristics of, and/or data stored on, remote systems using network services and/or agents; and/or the like.

[0106] Using these or any other suitable conditions, a control object can express rules that govern how content can be rendered, transferred, exported, and/or the like. It will be appreciated that the above list of conditions is illustrative in nature, and that any suitable conditions could be defined and used by, e.g., implementing a system call for use in testing for the desired condition. For example, without limitation, if it were desired to require that a device be located on a particular sub-network, a system call could be defined (e.g., GetIPConfig) that would be operable to return the host device's IPConfig information (or a remote device's IPConfig information, if the system call were run on a remote device using an agent), which could be used by a control program to test for whether the device was located on the prescribed sub-network.

[0107] 1.4.2. Agents

[0108] Preferred embodiments of the DRM engine-related systems and methods described herein provide support for independent objects that carry control programs. Such “agents” can be distributed to a DRM engine running on a remote system in order to accomplish specified functions, such as writing into the remote DRM engine's secure state store. For example, an agent could be sent as a consequence of contacting a remote service, or executing a remote control program. An agent can also be used to effect a content move operation, to initialize a counter, to deregister a node, and/or the like. As yet another example, an agent could be used to perform a reachability analysis from a remote node to another node. Such an agent could, e.g., be useful in enforcing a policy that prohibited a device registered to a first user from being registered to a second user. If the second user requested registration, an agent could be sent to the device by the second user, or a registration service acting on his or her behalf, to determine if the device was already registered to the first user, in which case the second user's registration request would be denied.

[0109] FIGS. 7A and 7B illustrate the use of agents in one embodiment. As shown in FIG. 7A, assume that two entities—system A 700 and system B 702—wish to communicate with each other over a computer network 703, and that a DRM system is being used that is capable of describing and enforcing rules for certain operations, such as accessing protected content, or creating DRM objects that can be used to represent memberships, registration status, and/or the like. In some cases, the rule(s) will be evaluated on system A 700, but will require information that depends on the state of system B 702. That information needs to be trusted by the DRM system 704 that is enforcing the rule(s) on system A 700.

[0110] For example, the DRM system 704 on system A 700 may be evaluating/enforcing a rule for performing a remote rendering of content from system A 700 to system B 702, and the rule might indicate that such an operation is permitted only if system B 702 is part of a certain group of devices, where the membership in that group is asserted by the presence of a state object 711 in a secure state database 716 accessible on system B 702.

[0111] A method used in a preferred embodiment to handle such situations makes use of agents. For example, if system A 700 needs information from system B 702, system A 700 prepares an agent 705, which, in one embodiment, is a control program (e.g., a sequence of instructions that can be executed by a DRM engine) that is sent from system A 700 to system B 702. In one embodiment, system A 700 sends agent code 705 to system B 702 over an authenticated communication channel 720 so that system A 700 can be confident that it is indeed on system B 702 that agent 705 will run. In some embodiments, along with agent code 705, system A 700 may also communicates to system B 702 one or more parameters that may be used by agent code 705 to perform its work.

[0112] As shown in FIG. 7B, system B 702 receives agent 705 and any associated agent parameters, and runs the agent code 705. When agent 705 is run on system B 702, it accesses system B's state database 716, retrieves state information 711 and/or performs one or more computations therewith, and sends the results 713 back to system A 700, preferably over authenticated communication channel 710. At this point, system A 700 has the information it needs to continue with its evaluation.

[0113] 1.4.3. Link Constraints

[0114] In one embodiment, the set of routines that represent the rules that govern the performance of a certain operation (such as “play”) on a content item is called an “action control”. The set of routines that represent validity constraints on a link object is called a “link constraint”. Like action controls, in preferred embodiments link constraints can express any suitable combination of conditions. Also like action controls, link constraints can be evaluated locally and/or remotely using a services interface or an agent.

[0115] 1.4.4. Obligations and Callbacks

[0116] In one embodiment, certain actions, when granted, require further participation from the host application. Obligations represent operations that need to be performed by the host application upon or after the use of the content key it is requesting. Callbacks represent calls to one or more of the control program's routines that need to be performed by the host application upon or after the use of the content key it is requesting. Examples of obligations include, without limitation, a requirement that certain outputs and/or controls be turned off while content is being rendered (e.g., to prevent writing the content to an unprotected output or to prevent fast-forwarding through certain important segments of the content); a requirement that information regarding use of the content be recorded (e.g., metering or audit information) and/or sent to a remote site (e.g., a clearinghouse, service provider, or the like); a requirement that an agent program be executed locally or remotely; and/or the like. Examples of callbacks include, without limitation a requirement that the host call the control program back at a certain absolute time, after a certain elapsed time (e.g., an elapsed time of content usage), after occurrence of a certain event (e.g., the completion of a trial content-rendering period), when the content has stopped being used, and/or the like. For example, a callback after a certain elapsed time could be used to increment or decrement budgets, playcounts, and the like (e.g., only debiting the users budget if they use a piece of content for at least a certain amount of time), thus protecting the user from having his or her account debited if he or she accidentally presses the play button but immediately presses stop.

[0117] In one embodiment, there are different types of obligations and callbacks, and if an application encounters any critical obligation or callback that it does not support, or does not understand (for example because the obligation type may have been defined after the application was implemented), the application is required to refuse to continue the action for which this obligation or callback parameter was returned.

[0118] 1.4.5. Example

[0119] FIGS. 8-12 show an example of how an illustrative embodiment of a DRM engine might control the use of a piece of content. Referring to FIG. 8, assume that the DRM engine has received a request to play a group 800 of content items 802, 804. For example, content items 802, 804 might comprise different sub-parts of a multimedia presentation, different tracks of an album, different pieces of content obtained from a subscription service, email attachments, or the like. The request may have been received by the DRM engine from a host application, which, in turn, received the request from a user of the computing device upon which the host application was running. The request from the host application will typically identify the requested action, the piece or pieces of content upon which the action is to be taken, and the license(s) that govern the content. DRM engine follows the process illustrated in FIG. 5 to determine whether the request should be granted.

[0120] FIGS. 8 and 9 provide a more detailed non-limiting example of the process shown in FIG. 5. Referring to FIG. 9, upon receiving the request to access content items 802 and 804 (block 900), the DRM engine examines the license(s) identified in the request, or otherwise in its possession, to see if a valid license exists. For example, the DRM engine might first identify the protector objects 806 and 808 that contain the unique identifiers of content items 802 and 804 (i.e., NS:007 and NS:008, respectively) (block 902 in FIG. 9). Next, the DRM engine locates the ContentKey objects 810 and 812 identified in protector objects 806 and 808 (block 904 in FIG. 9), which, in turn, enables the DRM engine to identify controller 814 which references both ContentKey objects 810 and 812 (block 906 in FIG. 9). In a preferred embodiment, controller 814 is signed, and DRM engine verifies its signature (or asks host services to verify it). The DRM engine uses controller 814 to identify the control object 816 that governs use of ContentKey objects 810 and 812 (and, thus, content items 802 and 804)(block 908 in FIG. 9). In a preferred embodiment, the DRM engine verifies the integrity of control object 816 (e.g., by computing a digest of control object 816 and comparing it to a digest contained in controller 814. If the integrity verification succeeds, the DRM engine executes the control code contained in control object 816 (block 910), and returns the result (block 912) to the host application, which uses it to grant or deny the user's request to access the content. The result of the control code might also optionally specify one or more obligations or callbacks which the host application will need to fulfill.

[0121] FIG. 10 is a more detailed example of how a DRM engine might perform the actions specified in blocks 910 and 912 of FIG. 9 (i.e., executing a control program and returning the result). As shown in FIG. 10, upon identifying the relevant control object, the DRM engine loads the byte code contained in the control object into a virtual machine that is preferably hosted by the DRM engine (block 1000). The DRM engine and/or the virtual machine will also typically initialize the virtual machine's runtime environment (block 1002). For example, the virtual machine might allocate the memory needed for execution of the control program, initialize registers and other environment variables, and/or obtain information about the host environment in which the virtual machine is operating (e.g., by making a System.Host.GetObject call, as described below). It will be appreciated that in some embodiments blocks 1000 and 1002 could effectively be combined or interleaved, and/or their order reversed. As shown in FIG. 10, the virtual machine next executes the control program's byte code (block 1004). As described elsewhere herein, this may involve making calls to other virtual machine code, retrieving state information from secure storage, and/or the like. When the control program has finished executing, it provides an output (e.g., in a preferred embodiment, an ExtendedStatusBlock) that may, for example, be used by the calling application to determine whether a request has been granted, and, if so, whether any obligations or callbacks are associated therewith; whether a request has been denied, and, if so, the reason for denial; or whether any errors occurred during execution (block 1006).

[0122] As previously indicated, the control code contained in control object 816 specifies the conditions or other requirements that must be satisfied in order to make the requested use of content items 802 and 804. The systems and methods described herein enable the specification of arbitrarily complex sets of conditions; however, for purposes of this example, assume that the control program is designed to require that, in order to play content items 802 and 804, (a) a given user's node must be reachable from the device on which the request to play the content was made, and (b) the current date must be after a specified date.

[0123] FIG. 11 shows how an illustrative embodiment of a DRM engine 1100 running on a device 1102 might execute the example control program described above, and FIG. 12 is a flowchart of the steps involved in the execution process. As shown in FIG. 11, DRM engine 1100 creates a virtual machine execution context (e.g., by calling System.Host.SpawnVm) 1104 and loads the control program. Virtual machine 1104 begins execution of the control program at the entry point specified by DRM engine 1100 (e.g., at the location of the Control.Actions.Play.perform routine). In this example, the control program needs to determine whether a given node is reachable from the personality node of the device 1102 on which the DRM engine 1100 is running. To make this determination, the control program makes a call 1105 to a link manager service 1106 provided by the DRM engine 1100, specifying the node to which linkage is required (block 1200 in FIG. 12). Link manager 1106 is responsible for evaluating link objects to determine if one node is reachable from another. To do this efficiently, link manager 1106 may pre-compute whether a path exists from the personality node 1110 of device 1102 to the various nodes 1114 specified in any link objects that device 1102 possesses. That is, link manager 1106 may, simply by checking the “to” and “from” fields of the links to which it as access, determine which nodes are potentially reachable from the personality node 1110 of device 1102. When link manager 1106 receives the call 1105 from virtual machine 1104, it determines whether the specified node 1112 is reachable by first determining if a path exists from personality node 1110 to the specified node 1112 (e.g., by checking for the node's ID in the list of nodes that it previously determined to be theoretically reachable)(block 1202 in FIG. 12). If a path exists, link manager 1106 evaluates any control programs contained in the links to see if the links are valid (blocks 1204-1210 in FIG. 12). To evaluate the control programs in the link objects (block 1206 in FIG. 12), link manager 1106 may use its own virtual machine 1108, on which it executes the control programs included in the link objects. Link manager 1106 returns the results of its determination (i.e., whether the given node is reachable) to the control program executing in virtual machine 1104, where it is used in the overall evaluation of whether the request to play the piece of content will be granted. Upon determining that the specified node 1112 is reachable from the personality node 1110 of device 1102, the control program executing on virtual machine 1104 next determines if the specified date restriction is met (block 1212 in FIG. 12). If the date restriction has been met (i.e., a “yes” exit from block 1212), then the control program returns a result indicating that the specified conditions have been met (block 1214 in FIG. 12); otherwise, control program returns a result indicating that the specified conditions were not satisfied (block 1216 in FIG. 12).

[0124] An example of a control program such as that described above is shown below:

;  Sample Control
;
;  This Control checks that a user node is reachable
;  and that the date is after a specific start date
;  and before a specific end date
;  The values are retrieved from attributes in the control
; =========================
; constants
; =========================
.equ DEBUG_PRINT_SYSCALL,	1
.equ FIND_SYSCALL_BY_NAME,	2
.equ SYSTEM_HOST_GET_OBJECT_SYSCALL,	3
.equ SUCCESS,	0
.equ FAILURE,	−1
; =========================
; data
; =========================
.data
ControlTargetNodeIdAttributePath:
.string “Octopus/Control/Attributes/TargetNodeId”
ControlStartDateAttributePath:
.string “Octopus/Control/Attributes/StartDate”
ControlEndDateAttributePath:
.string “Octopus/Control/Attributes/EndDate”
TargetNodeId:
.zeros 256
StartDate:
.long 0
EndDate:
.long −1
IsNodeReachableFunctionName:
.string “Octopus.Links.IsNodeReachable”
IsNodeReachableFunctionNumber:
.long 0
GetTimeStampFunctionName:
.string “System.Host.GetLocalTime”
GetTimeStampFunctionNumber:
.long 0
; =========================
; code
; =========================
.code
Global.OnLoad:
; load global functions
; get the syscall number for Octopus.Links.IsNodeReachable
PUSH @IsNodeReachableFunctionName
PUSH FIND_SYSCALL_BY_NAME
CALL
DUP
PUSH @IsNodeReachableFunctionNumber
POKE
BRN OnLoad_Fail
; get the syscall number for System.Host.GetTimeStamp
PUSH @GetTimeStampFunctionName
PUSH FIND_SYSCALL_BY_NAME
CALL
DUP
PUSH @GetTimeStampFunctionNumber
POKE
BRN OnLoad_Fail
; ok
PUSH 0
RET
OnLoad_Fail:
PUSH FAILURE
RET
Control.Actions.Play.Init:
; get the values from the attributes
; get the target node (guaranteed to be there)
PUSH 256	; ReturnBufferSize (256 bytes)
PUSH @TargetNodeId	; Return value
PUSH @ControlTargetNodeIdAttributePath	; Name
PUSH 0	; Parent = root container
PUSH SYSTEM_HOST_GET_OBJECT_SYSCALL
CALL
; get the start date
PUSH 4	; ReturnBufferSize (4 bytes)
PUSH @StartDate	; Return value
PUSH @ControlStartDateAttributePath	; Name
PUSH 0	; Parent = root container
PUSH SYSTEM_HOST_GET_OBJECT_SYSCALL
CALL
; get the end date
PUSH 4	; ReturnBufferSize (4 bytes)
PUSH @EndDate	; Return value
PUSH @ControlEndDateAttributePath	; Name
PUSH 0	; Parent = root container
PUSH SYSTEM_HOST_GET_OBJECT_SYSCALL
CALL
; success
PUSH 0
PUSH SUCCESS
STOP
Control.Actions.Play.Perform:
Control.Actions.Play.Check:
; check that the target node is reachable
PUSH @TargetNodeId
PUSH @IsNodeReachableFunctionNumber
PEEK
CALL
BRN Play_Fail
; put the current time on the stack
PUSH @GetTimeStampFunctionNumber
PEEK
CALL
; check that the date is before the end date
DUP   ; current time
PUSH @EndDate
PEEK
SWAP
CMP
BRN Play_Fail
; check that the date is after the start date
; the current time is on the stack
PUSH @StartDate
PEEK
CMP
BRN Play_Fail
; success
PUSH 0
PUSH SUCCESS
STOP
Play_Fail:
PUSH 0
PUSH FAILURE
STOP
.export Global.OnLoad
.export Control.Actions.Play.Init
.export Control.Actions.Play.Check
.export Control.Actions.Play.Perform

[0125] An additional example of a control program is included in Appendix E.

3. Content Consumption and Packaging Applications

[0126] The following is a more detailed description of illustrative embodiments of an application that consumes DRM-protected content (e.g., a media player, a word processor, an email client, etc., such as applications 303a, 303c, and 303d in FIG. 3), and a packaging application, such as application 303b, that packages content targeted to consuming applications.

[0127] 1.5. Content-Consuming Application Architecture

[0128] A content-consuming application will typically focus on accessing protected content, or could be part of a general-purpose application that also performs other functions, such as packaging content. In various embodiments, a content-consuming application might perform some or all of the following:

• • [0129] Provide an interface by which a user can request access to protected content objects and receive information about the content or error information;
  • [0130] Manage interaction with the file system;
  • [0131] Recognize the format of protected content objects;
  • [0132] Request a DRM engine to evaluate licenses for pieces of content to see if permission to access the content can be granted;
  • [0133] Verify digital signatures and deal with other general-purpose cryptographic functions that the DRM engine needs performed;
  • [0134] Request the DRM engine to provide the keys needed to decrypt protected content; and/or
  • [0135] Decrypt the protected content and interact with media rendering services to render the content.

[0136] In one embodiment, a DRM client engine evaluates the licenses associated with content, confirms or denies permission to use the content, and provides decryption keys to the content-consuming application. The DRM client engine may also issue one or more obligations and/or callbacks to the content-consuming application, requiring the application to perform certain actions as a consequence of having been given access to the content.

[0137] FIG. 13 shows the elements that make up a content-consuming client application 1300 in one embodiment. As shown in FIG. 13, host application 1302 is the logical central point of the client. It is responsible for driving the interaction pattern between the other modules, as well as interaction with the user through user interface 1304. The host application 1302 provides a set of services to DRM engine 1306 via a host services interface 1308. The host services interface 1308 allows the DRM engine 1306 to get access to data managed by the host application 1302, as well as certain library functions implemented by the host application 1302. In one embodiment, the host services interface 1308 it is the only outbound interface for the DRM engine 1306.

[0138] In one embodiment, the DRM engine 1306 does not interact directly with the multimedia content managed by the host application 1302. The host application 1302 logically interacts with content services 1310 for accessing the multimedia content, and passes on to the DRM engine 1306 only the portions of data that must be processed by the engine. Other interactions with the content are performed by the media rendering engine 1312. For example, in one embodiment content services 1310 are responsible for acquiring content from media servers, and storing and managing the content on the client's persistent storage, while media rendering engine 1312 is the subsystem responsible for accessing the multimedia content and rendering it (e.g., on a video and/or audio output). In one embodiment, the media rendering engine 1312 receives some information from DRM engine 1306 (such as content decryption keys), but in one embodiment the DRM engine 1306 does not interact with media rendering engine 1312 directly, but rather through the host application 1302.

[0139] Some of the information needed by the DRM engine 1306 might be available in-band with the multimedia content, and can be acquired and managed through the content services 1310, but some of this information may need to be obtained via means of other services such as a personalization service or a membership service (not shown).

[0140] In the embodiment shown in FIG. 13, cryptographic operations (e.g., encryption, signature verification, etc.) are handled by crypto services block 1314. In one embodiment, the DRM engine 1306 does not interact directly with the crypto services block 1314, but instead interacts indirectly via the host 1302 (using host services interface 1308), which forward its requests. Crypto services 1314 may also be used by, e.g., the media rendering engine 1312 in order to perform content decryption

[0141] It will be appreciated that FIG. 13 is provided for purposes of illustration, and that in other embodiments the various components shown in FIG. 13 could be rearranged, merged, separated, eliminated, and/or new components could be added. For example, without limitation, the logical division of functionality between the DRM engine and the host application in FIG. 13 is simply illustrative of one possible embodiment, and in practical implementations variations can be made. For example, the DRM engine could be integrated wholly or partially with the host application. Thus, it will be appreciated that any suitable division of functionality between host application and DRM engine can be used.

[0142] 1.6. Packager Architecture

[0143] The following provides an example of the functions that a packaging engine might perform for a host application that packages electronic content. In practice, a packaging application may focus on packaging specifically, or could be part of a general-purpose application operating at a user system that also accesses protected content (either packaged locally or elsewhere, e.g., on a network).

[0144] In various embodiments, a packaging host application might perform some or all of the following:

• • [0145] Provide a user interface by which content and license information can be specified;
  • [0146] Encrypt content;
  • [0147] Create the DRM objects that make up a license; and/or
  • [0148] Create a content object that contains or references the content and contains or references a license

[0149] FIG. 14 shows the elements that make up a packaging application 1400 in one embodiment. DRM packaging engine 1416 is responsible for packaging licenses such as those described herein (e.g., licenses comprising DRM objects such as controls, controllers, protectors, and the like). In some embodiments, DRM packaging engine 1416 may also associate metadata a license to explain, in human-readable form, what the license does.

[0150] In one embodiment, a host application 1402 provides a user interface 1404 and is responsible for obtaining information such as content references and the action(s) the user (typically a content owner or provider) wants to perform (e.g., to whom to bind content, what content-usage conditions to include in a license, etc). User interface 1404 can also display information about the packaging process, such as the text of the license issued and, if a failure occurs, the reason for the failure. In some embodiments, some information needed by the host application 1402 may require the use of other services, such as authentication or authorization services, and/or membership through a Service Access Point (SAP). Thus, in some embodiments the packaging application 1400 and/or the host application 1402 may need to implement some or all of the following:

• • [0151] Media format services 1406: In one embodiment, this element is responsible for managing media format operations such as transcoding and packaging. It is responsible as well for content encryption, which is achieved via content encryption services module 1408.
  • [0152] General-purpose cryptographic services 1410: In one embodiment, this element is responsible for issuing/verifying signatures, as well as encrypting/decrypting some data. Requests for such operations could be issued by the Service Access Point 1414 or by the DRM packaging engine 1416 via host services interface 1412.
  • [0153] Content encryption services 1408: In one embodiment, this module is logically separated from the general-purpose cryptographic services 1410 because it does not know about the application. It is driven by the media format services at content packaging time with a set of keys previously issued by the DRM packaging engine 1416.
    4. Key Derivation

[0154] The following describes a key derivation system that fits naturally with preferred embodiments of the DRM engine and system architecture described herein, and/or can be used in other contexts. Some of the examples in the following section are taken from a reference implementation of a preferred embodiment of this key derivation system known as “Scuba”. Additional embodiments are described in the '551 application.

[0155] As shown in FIG. 15, in some embodiments link objects 1530a, 1530b are used to distribute keys, in addition to their primary purpose of establishing relationships between nodes 1500a, 1500b, 1500c. As described above, a control object can contain a control program that can be used to decide if a request to perform an action should be granted or not. To do this, the control program may check whether a specific node is reachable via a chain of links. The key derivation techniques described herein take advantage of the existence of this chain of links to facilitate the distribution of a key, such that the key can be made available to the DRM engine that is executing the control program.

[0156] In one illustrative embodiment, each node object 1500a, 1500b, 1500c in a given deployment that uses the optional key distribution system has a set of keys that are used to encrypt content keys and other nodes' keys. Link objects 1530a, 1530b created for use in the same deployment contain some cryptographic data as a payload that allows key information do be derived when chains of links are processed by a DRM engine.

[0157] With nodes and links carrying keys in this manner, given a chain of links 1530a, 1530b from a node A 1500a to a node C 1500C, an entity (e.g., the DRM engine of a client host application) that has access to the secret sharing keys of node A 1515a, 1525a, also has access to the secret sharing keys of node C 1515c, 1525c. Having access to node C's secret sharing keys gives the entity access to any content key encrypted with those keys.

[0158] 1.7. Nodes, Entities, and Keys

[0159] 1.7.1. Entities

[0160] In one embodiment of a DRM system, nodes are data objects, not active participants in the system. Active participants, in this context, are called entities. Examples of entities are media players, devices, a subscription service, content packagers, and the like. Entities typically have nodes associated with them. An entity that consumes content uses a DRM engine and manages at least one node object that constitutes its personality. In one embodiment, an entity is assumed to have access to all the data of the node objects it manages, including all the private information of those objects.

[0161] 1.7.2. Nodes

[0162] Node objects that participate in an illustrative embodiment of the key derivation system contain keys as part of their data. In one embodiment, nodes may contain two general types of keys: sharing keys and confidentiality keys. The following sections list the different key types that can be used in various embodiments. It will be appreciated, however, that a specific deployment may use only a subset of these keys. For example, a system could be configured to work only with key pairs, omitting the use of secret symmetric keys. Or a system could be deployed without provisioning nodes with confidentiality keys if it only needed to use the sharing keys.

[0163] 1.7.2.1. Sharing Keys

[0164] Sharing keys are public/private key pairs and/or symmetric keys that are shared by a node N and all the nodes Px for which there exists a link from Px to N that contains key derivation extensions.

[0165] Sharing Public Key: Kpub-share[N] This is the public part of a pair of public/private keys for the public key cipher. This key typically comes with a certificate so that its credentials can be verified by entities that want to cryptographically bind confidential information to it.

[0166] Sharing Private Key: Kpriv-share[N] This is the private part of the public/private key pair. The entity that manages the node is responsible for ensuring that this private key is kept secret. For that reason, this private key will generally be stored and transported separately from the rest of the node information. This private key can be shared downstream with other nodes through the key derivation extensions of links.

[0167] Sharing Symmetric Key: Ks-share[N] This is a key that is used with a symmetric cipher. As with the private key, this key is confidential, and the entity that manages the node is responsible for keeping it secret. This secret key can be shared downstream with other nodes through the key derivation extensions of links.

[0168] 1.7.2.2. Confidentiality Keys

[0169] Confidentiality keys are key pairs and/or symmetric keys that are only known to the entity that manages the node to which they belong. The difference between these keys and the sharing keys described above is that they will not be shared with other nodes through the key derivation extensions in links.

[0170] Confidentiality Public Key: Kpub-conf[N] This is the public part of a pair of public/private keys for the public key cipher. This key typically comes with a certificate so that its credentials can be verified by entities that want to cryptographically bind confidential information to it.

[0171] Confidentiality Private Key: Kpriv-conf[N] This is the private part of the public/private key pair. The entity that manages the node is responsible for ensuring that this private key is kept secret. For that reason, this private key will generally be stored and transported separately from the rest of the node information.

[0172] Confidentiality Symmetric Key: Ks-conf[N] This is a key that is used with a symmetric cipher. As with the confidentiality private key, this key is kept secret.

[0173] 1.8. Cryptographic Elements

[0174] Preferred embodiments of the key derivation and distribution systems described herein can be implemented using a variety of different cryptographic algorithms, and are not restricted to any specific choice of cryptographic algorithm. Nevertheless, for a given deployment or profile, all participating entities will generally need to agree on a set of supported algorithms (where the term profile will generally refer to the specification of a set of actual technologies used in a particular implementation (e.g., an RSA for key derivation; XML for encoding objects; MP4 for the file format, etc.) and/or other representation of the semantic context that exists when objects are defined in a practical deployment).

[0175] In one embodiment, deployments include support for at least one public key cipher (such as RSA) and one symmetric key cipher (such as AES).

[0176] The following notation will be used when referring to cryptographic functions:

• • [0177] Ep(Kpub[N], M) means “the message, M, encrypted with the public key, Kpub, of node, N, using a public key cipher”
  • [0178] Dp(Kpriv[N], M) means “the message, M, decrypted with the private key, Kpriv, of node, N, using a public key cipher”
  • [0179] Es(Ks[N], M) means “the message, M, encrypted with the symmetric key, Ks, of node, N, using a symmetric key cipher”
  • [0180] Ds(Ks[N], M) means “the message, M, decrypted with the symmetric key, Ks, of node, N, using a symmetric key cipher”

[0181] 1.9. Targeting of Content Keys

[0182] In a preferred embodiment, two types of cryptographic targeting are used. Targeting a content key to a target node's sharing keys means making that key available to all entities that share the secret sharing keys of that target node. Targeting a content key to a node's confidentiality keys means making that key available only to the entity that manages that node. Targeting of a content key is done by encrypting the content key, CK, carried in a ContentKey object using one or both of the following methods:

• • [0183] Public Binding: Create a ContentKey object that contains Ep(Kpub[N], CK)
  • [0184] Symmetric Binding: Create a ContentKey object that contains Es(Ks[N], CK)

[0185] In a preferred embodiment, symmetric binding is used where possible, as it involves a less computationally intensive algorithm, and therefore makes it less onerous to the receiving entity. However, the entity (typically, a content packager) that creates the ContentKey object may not always have access to Ks[N]. If the packager does not have Ks[N], then it can use public binding, since Kpub[N] is not confidential information and therefore can be made available to entities that need to do public binding. Kpub[N] will usually be made available to entities that need to target content keys, accompanied by a certificate that can be inspected by the entity to decide whether Kpub[N] is indeed the key of a node that can be trusted to handle the content key in accordance with some agreed-upon policy (e.g., that the node corresponds to an entity running a DRM engine and host application that comply with the functional, operational, and security policies of the system).

[0186] 1.10. Derivation of Keys Using Links

[0187] To allow an entity to have access to the sharing keys of all the nodes reachable from its personality node, in one embodiment link objects contain an optional key extension payload. This key extension payload allows entities that have access to the private/secret keys of the link's “from” node to also have access to the private/secret sharing keys of the link's “to” node. In this way, an entity can decrypt any content key targeted to a node that is reachable from its personality node (if the targeting was done using the target node's sharing keys).

[0188] In one embodiment, when a DRM engine processes link objects, it processes the key extension payload of each link in order to update an internal chain of keys to which it has access. In one embodiment, the key extension payload of a link, L, from node, F, to node, T, comprises either:

• • [0189] Public derivation information: Ep(Kpub-share[F], {Ks-share[T], Kpriv-share[T]})
  • [0190] or
  • [0191] Symmetric derivation information: Es(Ks-share[F], {Ks-share[T], Kpriv-share[T]})

[0192] Where {Ks-share[T], Kpriv-share[T]} is a data structure containing Ks-share[T] and Kpriv-share[T].

[0193] The public derivation information is used to convey the secret sharing keys of node T, Ks-share[T] and Kpriv-share[T], to any entity that has access to the private sharing key of node F, Kpriv-share[F].

[0194] The symmetric derivation information is used to convey the secret sharing keys of node T, Ks-share[T] and Kpriv-share[T], to any entity that has access to the symmetric sharing key of node F, Ks-share[F].

[0195] As for targeting content keys to nodes, the preferred payload to include in a link is the symmetric derivation information. This is possible when the link creator has access to Ks-share[F]. If not, then the link creator will fall back to including the public derivation information as the payload for the link.

[0196] Assuming that the DRM engine processing a link already had Ks-share[F] and Kpriv-share[F] in its internal key chain, after processing the link, L[F→T], it will also have Ks-share[T] and Kpriv-share[T].

[0197] Since, in one embodiment, links can be processed in any order, the DRM engine may not be able to do the key derivation computations at the time a given link, L, is processed. This might be due to the fact that, at that time, the DRM engine's key chain might not yet contain the keys of the “from” node of that link. In this case, the link is remembered, and processed again when new information becomes available to the DRM engine, such as after processing a new link, P. If the “to” node of link P is the same as the “from” node of link L, and the “from” node of link P is a reachable node, then the “from” node of link L will also be reachable, and the key derivation step adds the private sharing keys of the “from” node of link L to the key chain.

5. Implementation Examples

[0198] Several examples are provided below to illustrate how various embodiments of the systems and methods described herein could be applied in practice. The systems and methods described herein can enable a wide range of rights management and other functionality, and thus it will be appreciated that the specific examples that are given here are not intended to be exhaustive, but are rather illustrative of the scope of the inventive body of work.

[0199] 1.11. Example: Users, PCs, and Devices

[0200] Assume that you want to implement a DRM system that ties the right to play content to a particular user, and you want to make it easy for the user to play content on all the playback devices that he or she owns. Assume that you decide that you are going to provide users with software that enables them to add playback devices as needed (e.g., mobile players). Also assume, however, that you want to set some policy to limit the number of general-purpose devices to which the user can transfer the content, so that the user does not have the ability to act as a distribution agency.

[0201] Based on these system requirements, it might, for example, make sense to tie the licenses you create to users, and to establish relationships between users and the devices that they use. Thus, in this example, you might first decide what kinds of nodes you need to establish the sorts of relationships that you require. For example, you might define the following types of nodes:

• • [0202] User (e.g., an individual who owns the rights to use the content)
  • [0203] PC (e.g., a software application, running on a personal computer, that can play content and specify additional playback devices)
  • [0204] Device (e.g., a portable content-rendering device)

[0205] Each node object can include a type attribute that indicates whether the object represents a user, a PC, or a device.

[0206] Say, for example, that you decide to restrict the maximum number of PC node objects that can be attached to any one user at a particular time to four (4). You decide there is no need to restrict the number of devices attached to the user as long as you provide restriction on the number of PCs. Based on this, a control program can be set up to allow access if a relationship can be established between the user node and the node that requests access. That node, then, could be either a PC or a device.

[0207] FIG. 16 shows a system designed to fulfill the foregoing requirements. Server 1600 assigns a user node object 1602a, 1602b to each new user 1604a, 1604b, and manages the ability of users 1604a, 1604b to associate devices 1606, 1608 and PCs 1610, 1612 therewith for the purpose of accessing protected content. When a user 1604a wishes to associate a new device 1606 with his or her user node 1602a, server 1600 determines whether the device 1606 already contains personalization information 1614, as might be the case if the device 1606 was personalized at the time manufacture. If the device does contain personalization information 1614, server 1600 uses that personalization information 1614 to create a link 1616 from the device 1606 to the user's node 1602a, and sends link 1616 to the user's device 1606. When user 1604a obtains protected content 1618 (e.g., from server 1600 or from some other content provider), that content 1618 is targeted to the user's node 1602a (e.g., by encrypting the content's decryption key with one of the secret sharing keys associated with the user's node 1602a) and a license 1619 is associated therewith specifying the conditions under which the content can be accessed. When user 1604a attempts to play content 1618 on device 1606, the DRM engine 1620 running on device 1606 evaluates the license 1619, which indicates that the content 1618 can be played as long as user node 1602a is reachable. DRM engine 1620 evaluates link 1616, which shows that user node 1602a is reachable from device 1606, and grants user 1604a's request to access content 1618, e.g., by authorizing decryption of the content decryption key contained within license 1619.

[0208] Since the content decryption key, in this example, is encrypted using a secret key associated with the user's node 1602a, this secret key will need to be obtained in order to decrypt the content decryption key. If the optional key derivation techniques described elsewhere herein have been used, the user node's key can be obtained simply by decrypting the key derivation information contained in link 1616 using one of device 1606's secret keys. The decrypted key derivation information will contain the key needed to decrypt the content decryption key contained in license 1619 (or information from which it can be derived or obtained).

[0209] Referring once again to FIG. 16, assume user 1604a wishes to associate a new PC 1610 with his or her user node 1602a. Server 1600 verifies that the maximum number of PCs have not already been associated with user node 1602a, and authorizes PC 1610 to be associated with user node 1602a. To perform the association, however, server 1600 needs to obtain personalization information from PC 1610 (e.g., cryptographic keys, a unique identifier, etc.). If, however, the PC 1610 has not been previously personalized (as might be the case if the user simply downloaded a copy of the PC software) server 1600 will perform the personalization process (e.g., by creating a PC node object using the bootstrap protocol described elsewhere herein) or direct the user to a service provider who can perform the personalization process. Upon completion of the personalization process, server 1600 can create a link 1624 from PC 1610 to user node 1602a and send the link to the PC 1610, which could continue to use it as long as it remained valid.

[0210] The user could request to add additional PCs later, and the server would enforce the policy that limits the number of PC node objects per user to 4 (typically it would also provide the ability for users to remove PCs from its active list as needed).

[0211] As yet another example, assume now that the service provider has decided that users should be able to play any content that they own on any device that they own. The service provider might also wish to allow the user's PC software to create links to each of his or her devices, rather than requiring the user to contact server 1600. In such an embodiment, when the user wished to play content on a new device, the user's PC software would access the new device's confidential personalization information and use it to create a new link for that device (e.g., a link from the new device to the user's node 1602a). If the device was not personalized, then the PC software might access a remote service, or direct the device to access the remote service, to perform the personalization process. The PC software would then send the link to the new device, at which point the new device would be able to play the content as long as it remained valid, since, in one embodiment, once a link object exists there is no need to create another one unless the link object expires or is otherwise invalidated.

[0212] In the examples shown above, content is targeted to the user. To do this, a packager application chooses a new ID for the content, or uses an existing one, creates an encryption key and associated ContentKey object, as well as a protector object to bind the content object and the ContentKey object. The packager then creates a control object with a control program (e.g., compiled in byte code executable by the DRM engine's virtual machine) that allows the “play” action to take place if and only if the user node is reachable from the PC or device node that is requesting the action. Typically, the control, controller, protector and ContentKey objects are embedded in the packaged content if appropriate, so that the PCs and devices do not have to obtain them separately.

[0213] In one embodiment, when a device or a PC wants to play content, it follows a process such as that previously described in connection with FIG. 9. That is, the DRM engine finds the protector object for the content ID of the content, then the ContentKey object referenced by that protector, then the controller object that references that ContentKey object, and finally the control object referenced by that controller. The DRM engine executes the control program of the control object, which checks whether or not the user node is reachable. If the device or PC node has the necessary link objects to verify that there exists a path between its node and the user node, then the condition is met and the control program allows the use of the key represented in the ContentKey object. The media rendering engine of the device or PC can then decrypt and play the content.

[0214] 1.12. Example: Temporary Login

[0215] FIG. 17 is another example of a potential application of the DRM systems and methods described herein. This example is similar to the example in the preceding section, except here the policy that governs creation of link objects between PC node objects and user node objects allows for a temporary login of no more than 12 hours, as long as the user does not already have a temporary login on another PC. This feature would allow a user 1700 to take his content 1702 to a friend's PC 1704, log in to that PC 1704 for a period of time, and play the content 1702 on the friend's PC 1704.

[0216] To accomplish this, a link object 1710 would be created with a limited validity period. In one embodiment, this could be done as follows:

[0217] For ease of explanation, assume that the DRM-enabled consuming software 1714 required to play the DRM-protected content 1702 is already present on the friend's PC 1704. The file containing the content 1702 and license 1708 is transferred to the friend's PC 1704. When the user tries to play the content 1702, the software 1714 recognizes that there is no valid link object linking the local PC node with the node of the user who owns the content. Software 1714 prompts the user for his credentials 1712 (this could be provided via a username/password, a mobile phone authentication protocol, a smartcard, or any authentication system allowed under the policy of the system) and communicates with a backend system 1706. The backend system 1706 checks the attributes of the user node object and PC node object for which the link is requested, and checks that there is no active temporary login link object still valid. If those conditions are met, the backend service 1706 creates a link object 1710 linking the friend's PC node object and user's node, with a validity period limited to the requested login duration (e.g., less than 12 hours, to comply with the policy in this example). Having the link object 1710 now enables the friend's PC 1704 to play the user's content 1702 until the link 1710 expires.

[0218] 1.13. Example: Enterprise Content Management

[0219] FIG. 18 shows the high-level architecture of an illustrative system 1800 for managing enterprise documents (e.g., email, word processing documents, presentation slides, instant messaging text, and/or the like). In the example shown in FIG. 18, a document editing application (e.g., a word processor) 1802, an email client 1804, and a directory server (e.g., an Active Directory server) 1806 make use of a digital rights management (DRM) plug-in 1808, a network service orchestration layer 1810, a registration service 1812, and a policy service 1816 to facilitate management of documents, email messages, and/or the like in accordance with policies. In a preferred embodiment, the DRM plug-in 1808, network service orchestration layer 1810, policy service 1816, and registration service 1812 are implemented using the DRM engine and service orchestration technologies described elsewhere herein and in the '551 application. For example, in one embodiment DRM plug-in 1808 may comprise an embodiment of the DRM engine described above. It will be appreciated that while FIG. 18 shows an embodiment in which existing applications such as word processor 1802 and email client 1804 are integrated with the DRM engine via a plugin that the applications can call, in other embodiments the DRM engine could be included as an integral part of either or both of the applications themselves. It will also be appreciated that the illustrative system shown in FIG. 18 can be implemented within a single enterprise or may span multiple enterprises.

[0220] In the illustration shown in FIG. 18, the directory server 1806 may, for example, contain user profiles and group definitions. For example, a group called “Special Projects Team” may be set up by a company's system administrator to identify the members of the company's Special Projects Team.

[0221] In one embodiment the directory server 1806 may comprise an Active Directory server running web services, such as those described in the '551 application (and implemented, e.g., with standard IIS based technologies on the Windows® platform), that issue nodes, links, and licenses to the people in the Special Projects Team group based on content that is accessed. If membership changes in the group, then new tokens may be issued. For revocation of rights, the directory server 1806 can run a security metadata service based on technology such as that described in the '551 application (occasionally referred to herein as “NEMO” technology). In some embodiments, the client can be required to have an to-date time value or notion of time (based on whatever freshness value the company chooses to define (e.g., 1 week, 1 day, 1 hour, every 5 minutes, etc.)) in order to use DRM licenses. For example, a token that the security metadata service provides might include a trusted and authenticable time value. In some embodiments, the client can identify user node IDs in security metadata service interactions. Security metadata can be evaluated directly in the context of license controls to determine if a user still has a given membership. Security metadata can also return agents that can determine if relationships such as being a member in the Special Projects Team are valid. Thus, in some embodiments it is possible to leverage a company's existing authorization and authentication infrastructure (e.g., the company's Active Directory server) with just the addition of a few well-defined web services.

[0222] FIG. 19 shows an example of how a system such as that shown in FIG. 18 can be used to manage access to or other use of a document. In this example, a particular employee (John) might frequently work on highly confidential strategic projects, and may have already installed the DRM plugin 1908 for his applications (e.g., a word processing program 1902, an email program 1904, a calendar program, a program or program suite that integrates such programs, and/or the like). At some point during the creation of his document, John accesses a “permissions” pull-down menu item that has been added to his application's toolbar (action 1913). A permissions dialog box appears which contacts his company's Active Directory Server 1906 for a directory of individuals and groups that have been set up on the system. He selects “Special Projects Team” from the list, and elects to give everyone on the team permission to view, edit, and print the document. Using the NEMO service orchestration technologies described in the '551 application, the DRM plugin 1908 contacts a NEMO-enabled Policy Service extension 1916 to the Active Directory 1906 and requests a copy of the Policy to use to protect files for the Special Projects Team (action 1914). When John saves the document, the DRM plugin automatically encrypts the file 1912, and creates a license object targeted and bound to the group known as “Special Projects Team” 1910. The license 1910 allows the file 1912 to be accessed (e.g., viewed, edited, printed, etc.) by any device that can produce a valid chain of links from its Device Node to the Special Projects Team Group Node.

[0223] John can access the document 1912 because his device has a link to John's User Node, and it also has a link from John's User Node to the “Special Projects Team” Group Node. Likewise, if he forwards this document to others, they can only access it if they also can produce a valid chain of links to the “Special Projects Team” Group Node (e.g., by requiring that the Special Projects Team Node be reachable by the device).

[0224] John might save the file (already protected) on his computer, and later attach it to an email message (action 1920). For example, he might open an old email to his boss (George), attach the file as he normally does, and send the message. As shown in FIG. 20, George also has the DRM plugin 2000 installed on his computer 2014. When he logged in to his computer 2014, the plugin 2000 opportunistically checked all of the groups that he has been added to (action 2006), and downloaded new, refreshed links for any that had expired (action 2012). If he had been added to “Special Projects Team” since his last login, his plugin 2000 would download a Link Object 2008 that links his User Node to the “Special Projects Team” Group Node. This Link 2008 signifies that User Node “George” is a member of the Group Node “Special Projects Team”. In this example, assume Link Object 2008 has an expiration date after which it will no longer be valid (e.g., 3 days).

[0225] As shown in FIG. 21, when George tries to open the document (actions 2130, 2132), the DRM plugin 2108 checks the embedded (or attached) license, and learns that the “Special Projects Team” node must be reachable. His plugin 2108 constructs (and validates) a chain of links 2120, 2122 from his computer's Device Node to the User Node “George”; and from User Node “George” to Group Node “Special Projects Team” (action 2134). Since the device has a valid chain of Links 2120, 2122, his plugin 2108 permits access to the file.

[0226] As described elsewhere herein, in some embodiments links can also carry a secure chain of keys. Thus, in some embodiments, by producing a chain of Links to the Special Projects Team Node, the plugin can not only prove that it is permitted to access the content, but also that it is capable of decrypting a chain of keys that enable it to decrypt the content.

[0227] If, for example, another employee (“Carol”) receives John's email accidentally, and attempts to open the document, her DRM plugin will retrieve the license bundled with the file and evaluate the terms of the license. Her PC has a link to her User Node “Carol”; but since she is not a member of the team, there is no Link from “Carol” to the “Special Projects Team” Group Node. Since “Special Projects Team” is not reachable, she is not permitted to access the file.

[0228] If Carol is eventually added to the group “Special Projects Team”. The next time her DRM plugin refreshes her memberships, it will detect this new group, and download a Link Object that links her User Node to the Special Projects Team Node. Her plugin now has all of the links it needs to construct a chain from her Device Node to her User Node to the Special Projects Team Node. The Special Projects Team node now “is reachable” and she can open any documents or emails that are targeted to the Special Projects Team—even those that were created before she joined the team.

[0229] Assume that a month later George moves on to a new role and is removed from the Special Projects Team Group in the Active Directory. The next time George logs in, his plugin does not receive a new, refreshed Link Object associating his User Node “George” to the “Special Projects Team”. When, weeks later, he tries to open John's file, his plugin attempts to construct a chain of links to the Special Projects Team. His PC still has a link to the User Node “George” (George's PC still belongs to him); but the Link from “George” to the “Special Projects Team” has expired. Since “Special Projects Team” is not reachable, he is not permitted to access the file.

[0230] Assume that the company has a policy that requires access to all confidential information to be logged. In one such an embodiment, the policy for the Special Projects Team dictates that all licenses that are created for this group also need to require collection and reporting of usage information to, e.g., a central repository. Thus, in this example, when evaluating (e.g., executing) the control program in the license, the plugin executes the requirement to log the access and does so. For example, activity of consequences can be logged in a local protected state database such as that described herein, and when network connectivity is re-established the relevant content can be reported via services previously described.

[0231] FIG. 22 shows another illustrative system 2200 for managing electronic content within an enterprise. In the example shown in FIG. 22 an LDAP server 2206 is used to manage user profiles, group definitions, and role assignments, and contains a group definition called “Special Projects Team”, and a role definition of “Attorney”.

[0232] Assume that John is an attorney and wishes to send an email with an attachment to other members of the Special Projects Team. When John installs the DRM plug-in 2208 for his applications, it also installs items to his email toolbar. At some point during his composition of the email message, John accesses “Set Permissions” from a pull-down menu that was added to his toolbar. The DRM plug-in 2208 contacts a Policy Service 2216 and displays a list of corporate messaging policies from which to choose. John selects “Special Project DRM Template” and the DRM plug-in 2208 uses the NEMO protocol to request and ensure the authenticity, integrity, and confidentiality of policy object that it receives. The policy describes how the licenses that use this template should be created, including how they should be targeted and bound.

[0233] When John hits “Send”, the DRM plugin 2208 encrypts the message and attachment, and generates the associated license(s). The license requires that in order to access the email or the attachment, either the Special Projects Team Group Node or the “Attorneys” Group Node must be reachable.

[0234] The license(s) are bundled with the encrypted message payload and encrypted attachment. The message is subsequently sent to a list of recipients using standard email functionality. Since the license rules and encryption are not dependent on the addressing of the email, the fact that an incorrect email recipient might be erroneously included does not put the contents of the email or attachment at risk.

[0235] Since such an unintended recipient will not have a valid Link Object linking his User Node to the Special Projects Team, he is not permitted to access the content if and when he attempts to do so. Furthermore, since his device does not have the necessary chain of Links (and the keys they contain), his device does not even have the capability to decrypt the content.

[0236] However, if the unintended recipient, in turn, forwards the same, unmodified email using standard email functionality to a member of the Special Projects Team. That member will have a Link Object that Links his User Node to the “Special Projects Team” Group Node, and will be able to access the email's contents.

[0237] Assume that another attorney (“Bill”) at the company has also received a Link Object that associates him with the “Special Projects Team” Group Node. Bill can also view the file. If he forwards the message to a paralegal (“Trent”), who is neither an attorney nor associated with the Special Projects Team, Trent will not have a Link Object that connects him with the “Special Projects Team” Group Node, and he will not be able to access the document.

[0238] If Trent is subsequently added to the Special Projects Team group in the LDAP directory 2206, he will be given the necessary Link Object(s) and will be able to access the previously forwarded email.

[0239] If, as previously discussed, the company has a policy indicating that a reporting requirement be included in all licenses, then, in one embodiment, whenever a control program within one of these licenses is executed (e.g., when someone attempts to access the file), a reporting event can be triggered. The reporting step can additionally include an indicator as to whether or not access was granted or denied—this is a matter of implementation choice. If such an indicator is used, a log can be maintained of the number of attempts to access a particular document, and status or other information on each (e.g., success, failure, etc.).

[0240] As yet another example, assume that one of the members (“Stephen”) of the Special Projects Team travels to another company to perform work on the special project. Before leaving for the other company, Stephen's email client downloads a local copy of all the email in is Inbox. The protected report attached to one of these emails also includes an embedded (or attached) license. This license object includes both the rules for accessing the content as well as an encrypted content key. The only “missing link” required to access the content is the necessary link objects to reach the “Special Projects Team” Group Node.

[0241] Since, in this example, the company's policy is to allow Link Objects to remain valid for 3 days, the Link Object that links Stephen's User Node to the Special Projects Team Node, will remain valid while he is traveling and disconnected. If he attempts to access the file while offline, the Special Projects Team Group Node will still be reachable, and he will be permitted to access the file.

[0242] If, however, Stephen stays offline for more than three days, the Link Object linking him to the Special Projects Team will expire. The Special Projects Team Group Node will then no longer be reachable, and he will not be permitted to access the file.

[0243] If Stephen eventually travels to a location where he can connect to the company's system (e.g., via VPN), his DRM plug-in will request refreshed copies of Link Objects for each of the groups to which he belongs. Since he is still part of the “Special Projects Team” group, he will receive a new link object from his User Node to the Special Projects Team Group Node. This link replaces the ‘old’ link which has expired and is no longer valid.

[0244] Since the “Special Projects Team” Node is now reachable using this new, refreshed Link, he is once again able to access the protected report. The new link object will be valid for a period of 3 days, after which it will also expire.

[0245] As yet another example, assume that a member (“Sally”) of the Special Projects Team wishes to communicate with another team member via an instant messenger, save a copy of the communication, and give it to another member of the team (e.g., via an email attachment, a diskette, a dongle, or the like). In this example, the instant messenger client (and, potentially any other messaging or communication products which the company offers its employees) is linked to a DRM plugin which, as in the previous examples, accesses the Policy “Special Project DRM Template” that dictates how licenses are to be targeted and bound. When Sally attempts to save her instant messaging conversation (e.g., by selecting “Save-As”), the plug-in chooses an encryption key (e.g., randomly) and packages (encrypts) the text of the conversation. Per company policy, the DRM plugin then generates a license object that is targeted and bound to the Special Projects Team Group Node.

[0246] The file containing the protected IM transcript is bundled with the license to access the transcript contents. As in the previous examples, the License contains both the rules that govern access to the content, as well as an encrypted copy of the content key. Sally can transfer this bundled file to an email, USB dongle, diskette, etc. using standard ‘drag and drop’ procedures, and send it to someone else. Provided that the recipient's device can produce valid links to the Special Project Group Node, access to the content is permitted and possible.

[0247] Assume that Sally gives the file to John, who is also a member of the Special Projects Team. If John has a recently-refreshed Link Object that identifies him as a member of the Special Projects Team, he will be able to access the file. Per the company's policy, this Link Object contains an expiration date that will cause it to expire in three days. Therefore, even if John remains disconnected, he will still have access as long as that link remains valid.

[0248] If, at some later time, John leaves the Special Projects Team for another job assignment, and finds the USB dongle from Sally in his bag and attempts to open the file using his desktop computer, the Link Object associating his User Node to the Special Projects Team will have expired. Since he is no longer part of the team, the DRM plugin on his device no longer can acquire new, refreshed links. Since the “Special Projects Team” Group Node is no longer reachable by his device, access is not permitted.

[0249] Figuring that his laptop has not been connected to the network since he changed jobs, he also tries to open the file with that device. Since the maximum allotted time has passed, that Link is also no longer valid. In some embodiments, each time he attempts to access the file, a report can be generated and queued to be sent to a central repository.

[0250] The central repository receives multiple reports of unsuccessful attempts to access the file and flags a manager via email. The manager reminds John that he is no longer permitted to access the confidential material and asks for all files to be destroyed (even though the system indicates that access has not been granted).

[0251] As yet another example, assume that a governmental agency or outside auditor wishes to investigate or audit the Special Projects Team's handling of confidential information. To support the investigation, the company wishes to demonstrate audit records for access to sensitive information related to the Special Project.

[0252] To this end, the company first scans all cleartext message archives for any messages related to the Special Project. To their relief, they discover that, in adherence to company policy, no employees sent messages discussing the Special Project without appropriate DRM protection (e.g. outside of the system).

[0253] The company then uses the DRM access records to produce an audit trail detailing who was given access to protected information, and when.

[0254] Per company procedure, when the Special Projects Team Group was established, it also included the Chief Compliance Officer (CCO) by default. A Link Object for Chief Compliance Officer was created and saved to the archive server, which allows him or her to review the contents of all messages if needed in the future.

[0255] In this example, the policy defined for the Special Projects Team indicated that all Licenses generated by the team must include the requirement to report any attempted access to the file, including the date and time, UserNode, and whether or not access was granted. These reports were saved in an access log on a central repository.

[0256] The CCO checks the access logs for all accesses associated with the Special Projects Team prior to the date when any leak or other irregularity was suspected to have occurred. The CCO also searches the email, IM, and network backup archives for all message traffic and system files on or before that date. Since each file has an attached license (with content key), and the CCO has the necessary Link Objects to satisfy the requirements of the License, he or she is permitted to access the contents of each and every message that was accessed prior to the time in question.

[0257] The access logs and unencrypted message contents are made fully available to the agency/auditor as part of the investigation.

[0258] In some embodiments the policy for the Special Projects Team could also have included the requirement to set an expiration date for the all licenses related to the Special Project. For example, if the company were only statutorily required to keep records of this nature for a period of 1 year, they could indicate in the policy that Licenses expire one year following date of issue. In that case, the company might only keep records as long as legally required to do so. Even the CCO would not have access after that time.

[0259] In the foregoing discussion, reference has occasionally been made to “targeting” and “binding”. In preferred embodiments, targeting and binding represent two different, yet closely related processes. In preferred embodiments, “binding” is primarily a cryptographic process, concerned with protecting the key that was used to encrypt the content. When a License is ‘bound’ to a Node (for example the “Special Projects Team” Node), it can mean, e.g., that the content key is encrypted with the public key associated with that Node. Thus, only devices that have access to the private key of the Node will have the necessary key to decrypt the content (and in preferred embodiments, the only way to get access to the private key of a Node is to decrypt a chain of Links to that Node); however, simply having the correct private key only indicates that the device has the capability to decrypt the content, if it is also permitted to do so.

[0260] In preferred embodiments, whether or not a device is permitted to access the content is determined by a Control Program within the License, and specifically, how it is “targeted”. “Targeting” refers to adding a requirement in the Control Program to specify that a particular node (or nodes) “are reachable” to perform a use of the content. In the examples shown above, the Control Program typically specifies that a particular Node “Special Projects Team” is reachable by the consuming device.

[0261] In some instances, it may be desirable to have licenses targeted to more than one Node, such as a new product development team at a company (“Company”) that is working with multiple suppliers to bid on components for a new top secret product. Assume that during the early stages of the project, Supplier A and Supplier B (competitors) both have links to “SecretProjectX”. Supplier A wants its ideas to be shared with all members of SecretProjectX, but does not want them to inadvertently leak to Supplier B. Supplier A can target these licenses such that: (“SecretProjectX is reachable”) AND (“Supplier A is reachable” or “Company is reachable”). If Company inadvertently shares this information to everyone in Secret Project X (including Supplier B), those at supplier B will not be permitted to look at it, limiting any non-disclosure risk to Company and eliminating the prospect of Supplier A losing its trade-secrets.

[0262] 1.14. Example: Healthcare Records

[0263] FIG. 23 illustrates how the systems and methods described herein could be applied to manage healthcare records. Assume that medical records have different levels of confidentiality, and that it is desirable to grant different access rights to different entities in the system (e.g., patients, doctors, insurance companies, and the like). For example, it may be desirable to permit some records to be viewed only by the patient, to permit some records to be viewed only by the patient's doctor, to permit some records to be viewable by the patient but only editable by the patient's doctor, to permit some records to be viewable by all doctors, to permit some records to be viewed by all insurance companies, to permit some records to be viewable only by the patient's insurance company, and/or the like.

[0264] As shown in FIG. 23, this healthcare ecosystem 2300 can be modeled using DRM objects like nodes and links, such as those describe elsewhere herein. For example, nodes could be assigned to the patient 2302, the patient's doctors 2304, the patient's insurance company 2306, the patient's devices (2308, 2310) a specific one of patient's doctors 2312, the doctor's computing devices 2314, 2316, the group of all doctors 2318, the group of doctors of a certain specialty 2320, a medical institution 2322, an insurance company 2324, the computing devices used by the insurance company 2326, the group of all insurance companies 2328, and the like.

[0265] Assume that the patient's doctor uses his or her PC to create a medical record regarding the patient. For example, the medical record may comprise a document template with a number of fields for his or her notes, diagnoses, prescription instructions, instructions for the patient and/or the like. The template may also allow the doctor to select the security policies for governing the document and/or the individual field thereof. For example, the doctor's application may present a set of standard security policy choices, and, upon obtaining the doctor's selection, may automatically generate a license based on those choices and associate with the protected (e.g., encrypted) content of the medical record.

[0266] For purposes of this example, assume the license grants viewing access to the patient, to all healthcare providers who treat the patient, and to all insurance companies that provide coverage for the patient. Further assume, for the sake of illustration, that the license grants editing rights only to cardiologists at medical institution x.

[0267] The packaging application accepts the doctor's policy specification input (which may simply comprise a mouse click on a standard template) and generates a license that includes a control program such as that shown below:

	Action.Edit.Perform( ) {
	if (IsNodeReachable(“MedicalFoundationX”) &&
	IsNodeReachable(“Cardiologist”)) {
	return new ESB(ACTION_GRANTED);
	} else {
	return new ESB(ACTION_DENIED);
	}
	}
	Action.View.Perform( ) {
	if (IsNodeReachable(“PatientY”) ||
	IsNodeReachable(“HCPsPatientY”) ||
	IsNodeReachable(“ICsPatientY”) {
	return new ESB(ACTION_GRANTED);
	} else if (EmergencyException == TRUE) {
	return new ESB(ACTION_GRANTED, new
	NotificationObligation( )); }
	else {
	return new ESB(ACTION_DENIED);
	}
	}

[0268] The medical record and its associated license might then be stored in a central database of medical records, a database operated by the particular medical foundation, and/or the like. If patient Y subsequently visits another healthcare provider, and authorizes that healthcare provider as one of his approved healthcare providers (e.g., by signing an authorization form), that healthcare provider will obtain a link to the patient y approved healthcare providers node, which the healthcare provider would store on his computer system. If that healthcare provider were to then obtain the medical record created by doctor x, he would be able to gain viewing access to that medical record, since patient y's approved healthcare provider node would be reachable from the new healthcare provider's computer system. If on the other hand, an unapproved healthcare provider were to obtain a copy of the (encrypted) medical record, he would be unable to access it since none of the required nodes (i.e., patient y's node, the node for all of patient y's approved healthcare providers, and the node for all of patient y's approved insurance companies) would be reachable from his computing system.

[0269] Note, however, that the example control program shown above includes an override feature that can be invoked, e.g., in emergencies if, for example, a healthcare provider needs to access the protected medical record, but is unable to satisfy the conditions of the control program (e.g., because the healthcare provider attempting to make emergency access to the medical record has not previously been registered as a healthcare provider of patient Y). Note also, however, that invocation of the emergency access exception will cause information to be automatically recorded regarding the invocation and/or other circumstances, and, in this example, will also cause a notification to be sent (e.g., to the patient's preferred healthcare provider—i.e., an entity explicitly authorized by the patient—and/or the patient himself). The association of such obligations with the emergency exception may discourage abuse of the exception, since a record of the abuse would exist.

[0270] It will be appreciated that this example program has been provided to facilitate explanation of certain embodiments of the systems and methods described herein. For example, whether a system includes support for emergency exceptions will typically depend on the requirements and desires of the system architect. Thus, for example, some embodiments may not support emergency exceptions, others may support emergency exceptions, but limit the class of entities who can invoke such exceptions to the class of “all doctors” (e.g., by requiring that the EmergencyException flag be set to “true” AND the All Doctors node be reachable), and others still may support emergency exceptions, but not associate mandatory obligations therewith (since inability to comply with the obligation would, in a preferred embodiment, render the content inaccessible), relying instead on non-technical, legal or institutional means for enforcement (e.g., by trusting healthcare providers not to abuse the ability to invoke the exception, and/or relying on industry certification and the legal system to prevent abuse).

[0271] Yet another variation that could be made to the examples provided above might be to require stronger proof that a doctor, or a specifically named doctor, was actually the one accessing a medical record, as opposed to someone else sitting at the computer that the doctor uses to access records (and thus a computer potentially containing the links necessary to satisfy a reachability analysis). Such a stronger form of authentication could be enforced in any suitable manner. For example, it could be wholly or partially enforced at the application or system level by protecting the doctor's computer and/or the software used to access medical records using passwords, dongles, biometric identification mechanisms, and/or the like. Alternatively, or in addition, the control programs associated with certain medical records could themselves include an obligation or condition require such stronger identification, such as checking for the presence of a dongle, requiring the host to obtain a password, and/or the like.

[0272] 1.15. Example: Subscriptions

[0273] FIG. 24 is an illustration of how the systems and methods presented herein could be used in the context of an electronic subscription service. Say, for example, that a user (Alice) wishes to obtain a subscription to jazz music from an Internet service provider (XYZ ISP). The Internet service provider may offer a variety of different subscription options, including a trial subscription that is free of charge, but can only be used to play subscription content five times before expiring (e.g., by playing one song five times, by playing five different songs once each, or the like). The trial subscription also will only make the content available in slightly degraded form (e.g., reduced fidelity or resolution). Alice uses her personal computer to access the service provider's Internet website, and opts for the trial subscription. The service provider then issues a link object 2400 and an agent 2401 and sends them to Alice's personal computer 2406. The agent 2401 is operable to initialize a state in Alice's secure state database that will be used to keep track of the number of times Alice has used trial content. The link 2400 is from Alice's ISP account node (Alice@XYZ_ISP) 2402 to subscription node 2404 and includes a control program that, when Alice requests to play a piece of content, checks the current value of the state variable set by the agent 2401 to see if additional plays are allowed.

[0274] When Alice downloads a piece of content to her PC and attempts to play it, the DRM engine on her PC evaluates the license associated with the content, which indicates that subscription node 2404 must be reachable in order to play the content. Alice had previously registered her PC with her ISP, at which time she received a link 2405 from her PC node 2406 to her account node 2402. The DRM engine thus possess link objects 2405, 2400 connecting PC node 2406 to subscription node 2404; however, before granting Alice's request to play the content, the DRM engine first determines whether the links are valid by executing any control programs that the links contain. When the control program in link 2400 is executed, the DRM engine checks the state database entry to determine if 5 plays have already been made, and, if they have not, grants her request to play the content, but also issues an obligation to the host application. The obligation requires the host to degrade the content before rendering. The host application determines that it is able to fulfill this obligation, and proceeds to render the content. In order to enable Alice to preview content before counting that content against her five free trial-offer plays, the control program might also include a callback that checks, e.g., 20 seconds after a request to play a piece of content has been granted, to see if the content is still being played. If the content is still being played, the play count is decremented, otherwise it is not. Thus, Alice can select from any of the content items offered by the subscription service, and play any five of them before her trial subscription expires.

[0275] Once Alice's trial subscription expires, Alice decides to purchase a full, monthly subscription which enables her to play as many content items as she wishes for a monthly fee. Alice use's her PC to sign up for the subscription, and receives a link 2410 from her account node 2402 to the subscription node 2404. The link includes a control program indicating that the link is only valid for one month (e.g., the control program checks an entry in the state database to see if one month has elapsed since the link was issued). This link 2410 is sent to Alice's PC, along with an agent program that is operable to initialize an appropriate entry in the state database of the PC's DRM engine indicating the date on which the link was issued. When Alice downloads a piece of content from the subscription service and attempts to play it, her PC's DRM engine determines that a path to the subscription node exists comprised of links 2405, 2410. The DRM engine executes any control programs contained in links 2405, 2410 to determine if the links are valid. If less than a month has elapsed since link 2410 was issued, the control program in link 2410 will return a result indicating that link 2410 is still valid, and Alice's request to play the piece of content. If Alice attempts to play a piece of content she previously obtained during her free trial period, the DRM engine on her PC will perform the same analysis and grant her request. Since the license associated with the piece of content obtained during the trial period indicates that if the TrialState variable in the secure database is not set, the only condition is that the subscription node must be reachable, Alice can now access that content once again since the subscription node is once again reachable from Alice's PC, this time via link 2410, not link 2400, which is no longer valid. Thus, Alice does not need to obtain a second copy of the content item to replace the copy she obtained during the free trial offer. Similarly, if Alice obtains a piece of subscription content from her friend, Bob, who is also a subscriber to the same service, Alice will, in this example, be able to play that content, too, since the content's license simply requires that the subscription node be reachable, not that it be reachable via Bob's PC or account.

[0276] It will be appreciated that the above examples are simply intended to illustrate some of the functionality that can be enabled by the systems and methods described herein, and is not intended to suggest that subscriptions must be implemented in precisely the manner described above. For example, in other embodiments, the license associated with a piece of subscription content might be bound to a user's node, rather than the subscription node, thus preventing two subscribers from sharing content like Bob and Alice were able to do in the example described above. It will be appreciated that many other variations to the above examples could be made.

[0277] The table below provides some illustrative pseudo-code for the agent, link, and license control programs in the example described above:

================================
The subscription trial gives you access to up to 5 pieces of
subscription content. The content will be marked as rendered only after
20 seconds of rendering. Content rendered in the context of the trial
will have to be degraded by the rendering applicatio.
The real subscription will be renewed every month and has no such
limitations on the number or quality of the renderings.
The code of the agent is as follows:
==================================
TrialAgent( ) {
SetObject(“TrialState”, 5);
}
==================================
The code of the control program in the trial link will be:
==================================
Control.Link.Constraint.Check( ) {
if (GetObject(“TrialState”, 5) > 0) {
return SUCCESS;
} else {
return FAILURE;
}
}
==================================
When Alice registers for real to the subscription service, she gets back
a link (from: Alice, to: Subscription) and an agent
The code of the agent is as follows:
==================================
RealSubscriptionAgent( ) {
// erase the TrialState if present
trialState = GetObject(“TrialState”);
if (trialState != NULL) {
SetObject(“TrialState”, NULL); // erase
}
}
==================================
The code of the link will be:
==================================
Control.Link.Constraint.Check( ) {
if (GetTrustedTime( ) < ExpirationDate) {
return SUCCESS;
} else {
return FAILURE;
}
}
==================================
The content licenses targeted to the subscription all have the same
control program:
==================================
Action.Play.Perform( ) {
// first check if the subscription node is reachable
if (!IsNodeReachable(“SubscriptionNode”)) {
return new ESB(ACTION_DENIED);
}
// now check if the TrialState is present
if (GetObject(“TrialState) != NULL) {
// we're in the trial mode: we need a callback and an obligation
return new ESB(ACTION_GRANTED,
new OnTimeElapsedCallback (20, DecrementCounter),
new DegradeRenderingObligation( ));
} else {
// we're in paid subscription mode: just return ACTION_GRANTED
return new ESB(ACTION_GRANTED);
}
}
// code of the callback function of OnTimeElapsed
DecrementCounter( ) {
SetObject(“TrialState”, GetObject(“TrialState”) − 1); }
==================================

[0278] Referring once again to FIG. 24, Alice also has an account 2420 with her mobile service provider, which remains valid as long as she remains connected to the network. Alice is not required to make a special payment for the subscription, in exchange for which she gets sent a link; instead renewal links 2424 are sent to her phone automatically when she connects to the network. These links enable her to access any of the content items or services offered by the mobile service provider, which have licenses that require only that the subscription node 2422 be reachable. If Alice changes mobile service providers, she will unable to access previously acquired content once her links 2424 expire.

[0279] FIG. 25 shows an example of how a service provider might interact with a home network domain 2500. In this example, devices are registered to a home network domain which enforces a policy that allows up to 5 devices to belong to the domain at any one time. Although the Smith family's cable service provider did not provide the domain manager software used to set up the home network domain 2500, cable service provider knows that the domain manager has been implemented by a certified provider of home network domain manager software, and thus trusts the domain manager software to operate as intended. As shown in FIG. 25, the Smith family connects Alice's phone and PC, Carl's PVR, and Joe's PSP to the domain 2500, resulting in links being issued from each of these devices to the domain node 2500. When new content is received, e.g., at the PVR, discovery services such as those described in the '551 application enable the other devices in the domain to automatically obtain the content and any necessary links. Links are issued from the domain node 2500 to the service provider account node 2502. Some of the cable service provider's content has a license with an obligation that fast forward and rewind must be disabled so that advertisements will be viewed. Carl's PVR and PC Alice's PC are able to enforce the obligation, and thus can play the content. Alice's mobile phone is unable to enforce the obligation and thus denies access to the content.

[0280] 1.16. Additional Examples: Content and Rights Sharing

[0281] As the preceding examples illustrate, embodiments of the systems and methods presented herein enable electronic content to be shared in natural ways. For example, the systems and methods described herein can be used to enable consumers to share entertainment content with their friends and family members, and/or enjoy it on all of their family's devices, while simultaneously protecting against wider, unauthorized distribution. For example, automated peer-to-peer discovery and notification services can be used, such that when one device obtains content or associated rights, other devices can automatically become aware of that content, thereby providing a virtual distributed library that can be automatically updated. For example, in one embodiment if one user obtains content or rights on a portable device at one location, then comes home, the user's family's devices can automatically discover and make use of those rights. Conversely, if a user obtains rights on a device on his or her home network, his or her portable devices can discover and carry away that content for use elsewhere. Preferred embodiments of the systems and methods described herein can be used to create services and rights objects that allow the above-described scenarios to be completely automated, using, for example, the service discovery and inspection techniques described in the '551 application. For example, the devices registered to a particular domain may provide services to each other (e.g., sharing of rights and content), and/or remote services can be invoked to facilitate local sharing of content. The systems and methods described enable the creation of DRM frameworks that are not focused on preventing the creation of copies per se, but rather are designed to work harmoniously with network technology to allow content to be shared, while protecting against consumers becoming illicit distributors of the content.

[0282] Preferred embodiments of the DRM systems and methods described herein also enable the determination of rights without the verbose types of rights expressions characteristic of some other DRM systems. Instead, preferred embodiments use a set of crafted rights objects that can interact contextually. These objects describe relationships and controls among entities such as users, devices, content, and groups thereof. For example, such contextual interactions might allow a device to determine that a given piece of content can be played because (a) the content was obtained from a legitimate content service that the user currently subscribes to, (b) the user is part of a specific family group, and (c) the device is associated with this specific family group. There are numerous types of relationships such as those described in this example, which users understand intuitively, and preferred embodiments of the systems and methods described herein enable the creation of systems that naturally understand these kinds of relationships. The relationships among entities can be created, destroyed, and changed over time, and preferred embodiments provide a natural way of determining rights in a dynamic networked environment—an environment that consumers can naturally understand. Nevertheless, if a content deployer wants to use a more traditional rights expression approach, preferred embodiments can accommodate that as well. For example, tools can be used to translate traditional rights expressions into sets of objects such as those described above, and/or a DRM engine can be implemented that operates directly on such rights expressions. Alternatively, in some embodiments, devices do not need to understand such traditional rights expressions, and are not constrained by their limitations.

[0283] Preferred embodiments of the systems and methods described herein also have a very general notion of a media service. A broadcast service and an Internet download or subscription service are examples of media services. Restrictions associated with these services can make content difficult to share. With preferred embodiments of the systems and methods described herein, content can be obtained on broadcast, broadband, and mobile services, and shared on a group of networked devices in the home, including portable devices. Alternatively, or in addition, services can be offered by individual devices in a peer-to-peer fashion via wireless connectivity. For example, the new generation of WiFi enabled cellphones can provide content catalog services to other devices. Such a service allows other devices to “see” what content is available to be shared from the device. The service provides information that can be used to determine the rights so that any limitations can be accepted or easily eliminated.

[0284] Preferred embodiments of the systems and methods described herein are not confined to one service or to one platform. As explained above, preferred embodiments are capable of working with numerous services, including “personal” services. This is becoming more and more important as home and personal networks become more ubiquitous. For example, digital cameras are now available with WiFi connectivity, making it very convenient to share photos over networks. It is nice to be able to automate the sharing of photographs, but the camera will encounter many different networks as it is carried about. Automated sharing is convenient, but personal photos are, of course, personal. Embodiments of the systems and methods described herein make it easy to share photos within a family on the family's devices, but not with arbitrary devices that happen to encounter the camera on a network. In general, as more devices become networked, it is going to be increasingly important to manage the rights of all content on those devices. Although the purpose of networking is to allow information on the networked devices to be shared, networks will overlap and merge into one another. Networks enable content to be shared easily but it should not be shared arbitrarily. Thus, it is desirable to have a DRM system that is network-aware and that can use the context provided by the content, the user, the network, and characteristics of devices to determine if and how content should be shared. Preferred embodiments of systems and methods described herein enable such an approach.

6. Reference Architecture for Content Consumption and Packaging

[0285] The following is a description of a reference architecture for a consuming application (e.g., a media player) that consumes DRM-protected content, and a packaging application (e.g., an application residing on a server) that packages content targeted to consuming applications.

[0286] 1.17. Client Architecture

[0287] The following provides an example of functions that an illustrative embodiment of a DRM engine might perform for a host application that consumes content.

[0288] 1.17.1. Host Application to DRM Engine Interface

[0289] Although in a preferred embodiment there is no required API for DRM engines, the following are high-level descriptions of the type of interface provided by an illustrative DRM engine (referred to as the “Octopus” DRM engine) to a host application in one illustrative embodiment:

[0290] Octopus::CreateSession(hostContextObject)→Session—Creates a session given a Host Application Context. The context object is used by the Octopus DRM engine to make callbacks into the application.

[0291] Session::ProcessObject(drmObject)—This function should be called by the host application when it encounters certain types of objects in the media files that can be identified as belonging to the DRM subsystem. Such objects include content control programs, membership tokens, etc. The syntax and semantics of those objects is opaque to the host application.

[0292] Session::OpenContent(contentReference)→Content—The host application calls this function when it needs to interact with a multimedia content file. The DRM engine returns a Content object that can be used subsequently for retrieving DRM information about the content, and interacting with it.

[0293] Content::GetDrmInfo( )—Returns DRM metadata about the content that is otherwise not available in the regular metadata for the file.

[0294] Content::CreateAction(actionInfo)→Action—The host application calls this function when it wants to interact with a Content object. The actionInfo parameter specifies what type of action the application needs to perform (e.g., Play), as well as any associated parameters, if necessary. The function returns an Action object that can then be used to perform the action and retrieve the content key.

[0295] Action::GetKeyInfo( )—Returns information that is necessary for the decryption subsystem to decrypt the content.

[0296] Action::Check( )—Checks whether the DRM subsystem will authorize the performance of this action (i.e whether Action::Perform( ) would succeed).

[0297] Action::Perform( )—Performs the action, and carries out any consequences (with their side effects) as specified by the rule that governs this action.

[0298] 1.17.2. DRM Engine to Host Services Interface

[0299] The following is an example of the type of Host Services interface needed by an illustrative embodiment of a DRM engine from an illustrative embodiment of a host application.

[0300] HostContext::GetFileSystem(type)→FileSystem—Returns a virtual FileSystem object that the DRM subsystem has exclusive access to. This virtual FileSystem will be used to store DRM state information. The data within this FileSystem should only be readable and writeable by the DRM subsystem.

[0301] HostContext::GetCurrentTime( )—Returns the current date/time as maintained by the host system.

[0302] HostContext::GetIdentity( )—Returns the unique ID of this host.

[0303] HostContext::ProcessObject(dataObject)—Gives back to the host services a data object that may have been embedded inside a DRM object, but that the DRM subsystem has identified as being managed by the host (e.g., certificates).

[0304] HostContext::VerifySignature(signatureInfo)—Checks the validity of a digital signature over a data object. In one embodiment the signatureInfo object contains information equivalent to the information found in an XMLSig element. The Host Services are responsible for managing the keys and key certificates necessary to validate the signature.

[0305] HostContext::CreateCipher(cipherType, keyInfo)→Cipher—Creates a Cipher object that the DRM subsystem can use to encrypt and decrypt data. A minimal set of cipher types will be defined, and for each a format for describing the key info required by the cipher implementation.

[0306] Cipher::Encrypt(data)

[0307] Cipher::Decrypt(data)

[0308] HostContext::CreateDigester(digesterType)→Digester—Creates a Digester object that the DRM subsystem can use to compute a secure hash over some data. In one embodiment, a minimal set of digest types can be defined.

[0309] Digester::Update(data)

[0310] Digester::GetDigest( )

[0311] 1.17.3. UML Sequence Diagram

[0312] FIG. 26 illustrates the use of the illustrative APIs set forth in the preceding sections, and the interactions that take place between the host application and the DRM client engine in an exemplary embodiment.

[0313] 1.18. Packager Reference Architecture

[0314] The following provides an example of the functions that a packaging engine might perform for a host application that packages content. In practice, a packaging application may focus on packaging specifically, or could be part of a general purpose application operating at a user system that also accesses protected content (either packaged locally or elsewhere in a network).

[0315] 1.18.1. Host Application to Packaging Engine Interface

[0316] This section provides a high-level description of an illustrative API between a host application and a packaging engine used in connection with a reference DRM engine referred to as “Octopus”.

[0317] Octopus::CreateSession(hostContextObject)→Session. Creates a session given a host application context. The context object that is returned by this function is used by the packaging engine to make callbacks into the application.

[0318] Session::CreateContent(contentReferences[ ])→Content. The host application calls this function in order to create a content object that will be associated with license objects in subsequent steps. Having more than one content reference in the contentReferences array implies that these are bound together in a bundle (e.g., one audio and one video track) and that the license issued should be targeted to these as one indivisible group.

[0319] Content::SetDrmInfo(drmInfo). The drmInfo parameter specifies the metadata of the license that will be issued. The drmInfo will act as a guideline to translate the license into bytecode for the virtual machine.

[0320] Content::GetDRMObjects(format)→drmObjects. This function is called when the host application is ready to get the drmObjects that the packager engine created. The format parameter will indicate the format expected for these objects (e.g., XML or binary atoms).

[0321] Content::GetKeys( )→keys[ ]. This function is called by the host packaging application when it needs keys in order to encrypt content. In one embodiment, there is one key per content reference.

[0322] 1.18.2. Packaging Engine to Host Services Interface

[0323] The following is an example of the type of interface that an illustrative packaging engine needs the host application to provide in one embodiment.

[0324] HostContext::GetFileSystem(type)→FileSystem. Returns a virtual FileSystem object that the DRM subsystem has exclusive access to. This virtual FileSystem can be used to store DRM state information. The data within this FileSystem should only be readable and writeable by the DRM subsystem.

[0325] HostContext::GetCurrentTime( )→Time. Returns the current date/time as maintained by the host system.

[0326] HostContext::GetIdentity( )→ID. Returns the unique ID of this host.

[0327] HostContext::PerformSignature(signatureInfo, data). Some DRM objects created by the packaging engine will have to be trusted. This service provided by the host will be used to sign a specified object.

[0328] HostContext::CreateCipher(cipherType, keyInfo)→Cipher. Creates a cipher object (an object that is able to encrypt and decrypt data) that the packaging engine can use to encrypt and decrypt data. In one embodiment, the cipher object is used to encrypt the content key data in the ContentKey object.

[0329] Cipher::Encrypt(data). Encrypts data.

[0330] Cipher::Decrypt(data). Decrypts data.

[0331] HostContext::CreateDigester(digesterType)→Digester. Creates a digester object that the packaging engine can use to compute a secure hash over some data.

[0332] Digester::Update(data). Feeds data to the digester object.

[0333] Digester::GetDigest( ). Computes a digest.

[0334] HostContext::GenerateRandomNumber( ). Generates a random number that can be used for generating a key.

[0335] FIG. 27 is a UML diagram showing an example of the use of the illustrative APIs set forth above, and the interactions that take place between the host application and the packaging engine in one illustrative embodiment.

7. Objects

[0336] This section provides more information regarding the DRM objects that serve as the building blocks of an illustrative implementation of a DRM engine. First, a relatively high-level overview is given of the types of objects the DRM engine uses for content protection and governance. Next, a more detailed description of these objects and the information they convey is provided, along with some example data structures used in one illustrative embodiment.

[0337] 1.19. Content Protection and Governance DRM Objects

[0338] As previously described in connection with FIG. 6, content governance objects (sometimes referred to, collectively with node and link objects, as “DRM objects”) are used to associate usage rules and conditions with protected content. Together, these objects form a license.

[0339] As shown in FIG. 6, the data represented by content object 614 is encrypted using a key. That key needed to decrypt the content is represented by ContentKey object 602, and the binding between the content and the key used to encrypt it is represented by protector object 604. The rules that govern the use of the decryption key are represented by control object 608, and the binding between the ContentKey 602 and the control object 608 is represented by controller object 606. In one embodiment, trusted systems will only make use of the content decryption key under governance of the rules expressed by the byte code in control object 608. FIG. 28A is a more detailed illustration of a license such as that shown in FIG. 6, and illustrates a signature scheme that is used in one embodiment.

[0340] 1.19.1. Common Elements

[0341] In one embodiment, objects share common basic traits: they can each have an ID, a list of attributes, and a list of extensions.

[0342] 1.19.1.1. IDs

[0343] Objects that are referenced by other objects have a unique ID. In one embodiment, IDs are simply URIs, and the convention is that those URIs are URNs

[0344] 1.19.1.2. Attributes

[0345] Attributes are typed values. Attributes can be named or unnamed. The name of a named attribute is a simple string or URI. The value of an attribute is of a simple type (string, integer, or byte array) or a compound type (list and array). Attributes of type ‘list’ contain an unordered list of named attributes. Attributes of type ‘array’ contain an ordered array of unnamed attributes.

[0346] An object's ‘attributes’ field is a (possibly empty) unordered collection of named attributes.

[0347] 1.19.1.3. Extensions

[0348] Extensions are elements that can be added to objects to carry optional or mandatory extra data. Extensions are typed, and have unique IDs as well. Extensions can be internal or external.

[0349] 1.19.1.3.1. Internal Extensions

[0350] Internal extensions are contained in the object they extend. They have a ‘critical’ flag that indicates whether the specific extension data type for the extension is required to be known to the implementation that uses the object. In one embodiment, if an implementation encounters an object with a critical extension with a data type that it does not understand, it must reject the entire object.

[0351] In one embodiment, the ID of an internal extension needs to be locally unique: an object cannot contain two extensions with the same ID, but it is possible that two different objects each contain an extension with the same ID as that of an extension of the other object.

[0352] An object's ‘extensions’ field is a (possibly empty) unordered collection of internal extensions.

[0353] 1.19.1.3.2. External Extensions

[0354] External extensions are not contained in the object they extend. They appear independently of the object, and have a ‘subject’ field that contains the ID of the object they extend. In one embodiment, the ID of an external extension needs to be globally unique.

[0355] 1.19.2. Content

[0356] In one embodiment, the content object is an “external” object. Its format and storage are not under the control of the DRM engine, but under the content management subsystem of the host application (for instance, the content could be an MP4 movie file, an MP3 music track, etc.). In one embodiment, the format for the content needs to provide support for associating an ID with the content payload data. The content payload is encrypted in a format-dependent manner (typically with a symmetric cipher, such as AES).

[0357] 1.19.3. ContentKey

[0358] The ContentKey object represents a unique encryption key, and associates an ID with it. The purpose of the ID is to enable Protector objects and Controller objects to make references to ContentKey objects. The actual key data encapsulated in the ContentKey object is itself encrypted so that it can only be read by the recipients that are authorized to decrypt the content. The ContentKey object specifies which cryptosystem was used to encrypt the key data. The cryptosystem used to protect the content key data is called the Key Distribution System. Different Key Distribution Systems can be used. An example of a Key Distribution System is the Scuba Key Distribution System described above.

[0359] 1.19.4. Protector

[0360] The Protector object contains the information that makes it possible to find out which key was used to encrypt the data of Content objects. It also contains information about which encryption algorithm was used to encrypt that data. In one embodiment, the Protector object contains one or more IDs that are references to Content objects, and exactly one ID that is a reference to the ContentKey object that represents the key that was used to encrypt the data. If the Protector points to more than one Content object, all those Content objects represent data that has been encrypted using the same encryption algorithm and the same key. In one embodiment, unless the cryptosystem used allows a safe way of using the same key for different data items, it is not recommended that a Protector object point to more than one Content object.

[0361] 1.19.5. Control

[0362] The control object contains the information that allows the DRM engine to make decisions regarding whether certain actions on the content should be permitted when requested by the host application. In one embodiment, the rules that govern the use of content keys are encoded in the control object as byte code for execution by the virtual machine. The control object also has a unique ID so that it can be referenced by a controller object. In one embodiment, control objects are signed, so that the DRM engine can verify that the control byte code is valid and trusted before it is used to make decisions. The validity of the control object can also optionally be derived through the verification of a secure hash contained in a controller object.

[0363] 1.19.6. Controller

[0364] The controller object contains the information that allows the DRM engine to find out which control governs the use of one or more keys represented by ContentKey objects. The controller object contains information that binds it to the ContentKey objects and the control object that it references. In one embodiment, controller objects are signed (e.g., by a packager application that has a certificate allowing it to sign controller objects), so that the validity of the binding between the ContentKey and the control object that governs it, as well as the validity of the binding between the ContentKey ID and the actual key data, can be established. The signature of the controller object can be a public key signature or a symmetric key signature, or a combination of both. Also, when the digest of the control object referenced by the controller object is included in the controller object, the validity of the control object can be derived without having to separately verify the signature of the control object.

[0365] 1.19.6.1. Symmetric Key Signature

[0366] In one embodiment, this is the preferred type of signature for controller objects, and is implemented by computing a Message Authentication Code (MAC) of the controller object, keyed with the same key as the key represented by the corresponding ContentKey object. In one embodiment, the canonical method for this MAC is to use HMAC with the same hashing algorithm as the one chosen for the PKI signature algorithm used in the same deployment.

[0367] 1.19.6.2. Public Key Signature

[0368] This type of signature is used when the identity of the signer of the controller object needs to be known. This type of signature is implemented with a public key signature algorithm, signing with the private key of the principal who is asserting the validity of this object. In one embodiment, when using this type of signature, a symmetric key signature will also be present, and sign both the controller object as well as the public key signature, so that is can be guaranteed that the principal who signed with its private key also had knowledge of the actual value of the content key carried in the ContentKey object.

[0369] 1.20. Identity and Key Management DRM Objects

[0370] As previously described, node objects represent entities in a DRM profile, and no implicit or explicit semantics are used to define what the node objects represent. A given deployment (DRM profile) of a system will define what types of principals exist, and what roles and identities different node objects represent. That semantic information is typically expressed using attributes of the node object.

[0371] Link objects represent relationships between nodes. Link objects can also optionally contain some cryptographic data that allows the link to be used for content key derivation. Just as for nodes, in one embodiment no implicit or explicit semantics are used to define what a link relationship means. Depending on what the from and to nodes of the link represent in a given DRM Profile, the meaning of the link relationship can express membership, ownership, association, and/or many other types of relationships. In a typical DRM profile, some node objects could represent users, other nodes could represent devices, and other nodes could represent user groups or authorized domains (ADs). In such a context, links between devices and users might represent an ownership relationship, and links between users and user groups or authorization domains might represent membership relationships. FIG. 28B illustrates the structure and interrelationship between nodes and links in one example embodiment.

[0372] 1.20.1. Node

[0373] The node object represents an entity in the system. The node object's attributes define certain aspects of what the node object represents, such as the role or identity represented by the node object in the context of a DRM profile. The node object may also have a confidentiality asymmetric key pair that is used for targeting confidential information to the subsystems that have access to the confidential parts of the node object (typically, the entity represented by the node, or some entity that is responsible for managing that node). Confidential information targeted at a node can be encrypted with that node's confidentiality public key. The node object may also have a sharing asymmetric key pair and a sharing symmetric key can be used in conjunction with link objects when the system uses a ContentKey derivation system for ContentKey distribution, such as that described elsewhere herein. In a preferred embodiment, only entities that need to be referenced by link or control objects, or to receive cryptographically targeted information, need to have corresponding node objects.

[0374] 1.20.2. Link

[0375] The link object is a signed assertion that there exists a directed edge in the graph whose vertices are the node objects. For a given set of nodes and links, we say that there is a path between a node X and a node Y if there exists a directed path between the node X vertex and the node Y vertex in the graph. When there is a path between node X and node Y, we say that node Y is reachable from node X. The assertions represented by link objects are used to express which nodes are reachable from other nodes. The controls that govern content objects can check, before they allow an action to be performed, that certain nodes are reachable from the node associated with the entity performing the action. For example, if node D represents a device that wants to perform the “play” action on a content object, a control that governs the content object can test if a certain node, U, representing a certain user, is reachable from node D. To determine if node U is reachable, the DRM engine can check whether there is a set of link objects that can establish a path between node D and node U.

[0376] In one embodiment, the DRM engine verifies link objects before it uses them to decide the existence of paths in the node graph. Depending on the specific features of the certificate system (e.g., x509v3) used to sign link objects, link objects can be given limited lifetimes, be revoked, etc. In one embodiment, the policies that govern which keys can sign link objects, which link objects can be created, and the lifetime of link objects are not directly handled by the DRM engine. Instead, those policies leverage the node's attribute information. To facilitate the task of enforcing certain policies, in one embodiment, a way to extend standard certificate formats with additional constraint checking is provided. These extensions make it possible to express validity constraints on certificates for keys that sign links, such that constraints such as what type of nodes the link is connecting, as well as other attributes, can be checked before a link is considered valid.

[0377] In one embodiment, a link object can contain a control object that will be used to constrain the validity of the link. In addition, in one embodiment a link object may contain cryptographic key derivation data that provides the user with sharing keys for key distribution. That cryptographic data will contain, in addition to metadata, the private and/or symmetric sharing keys of the “from” node, encrypted with the sharing public key and/or the sharing symmetric key of the “to” node.

[0378] 1.21. Data Structures

[0379] The following paragraphs describe, in more detail, an illustrative object model for the objects discussed above, defining the fields that each type of object has in one illustrative embodiment. Data structures are described using a relatively simple object description syntax. Each object type is defined by a class that can extend a parent class (this is an “is-a” relationship). The class descriptions are in terms of the simple abstract types “string” (character strings), “int” (integer value), “byte” (8-bit value), and “boolean” (true or false) but do not define any specific encoding for those data types, or for compound structures containing those types. The way objects are encoded, or represented, can vary depending on the implementation of the engine. In practice, a given profile of use of the DRM engine can specify how the fields are represented (e.g., using an XML schema).

[0380] In one illustrative embodiment, the following notations are used:

class ClassName {	Defines a class type. A class type is a
field1;	heterogeneous compound data type
field2;	(also called object type). This
....	compound type is made up of one or
}	more fields, each of a simple or
	compound type. Each field can be of a
	different type.
type[ ]	Defines a homogeneous compound data
	type (also called list or array type). This
	compound type is made up of 0 or more
	elements of the same type (0 when the
	list is empty).
String	Simple type: represents a character
	string
Int	Simple type: represents an integer value
Byte	Simple type: represents an integer value
	between 0 and 255
Boolean	Simple type: represents a boolean value
	(true or false)
class SubClass extends	Defines a class type that extends
SuperClass {...}	another class type. A class that extends
	another one contains all the fields of the
	class it extends (called the superclass)
	in addition to its own fields.
Abstract class {...}	Defines an abstract class type. Abstract
	class types are types that can be
	extended, but are never used by
	themselves.
{type field;}	Defines an optional field. An optional
	field is a field that may be omitted from
	the compound data type that contains it.
(type field;)	Defines a field that will be skipped
	when computing the canonical byte
	sequence for the enclosing compound
	field
class SubClass extends	Defines a subclass of a class type and
SuperClass (field=value) {...}	specifies that for all instances of that
	subclass, the value of a certain field of
	the superclass is always equal to a fixed
	value.

[0381] 1.21.1. Common Structures

[0382] In one illustrative embodiment, the following common structures are used:

	abstract class Octobject {
	{string id;}
	Attribute[ ] attributes;
	InternalExtension[ ] extensions;
	}
	class Transform {
	string algorithm;
	}
	class Digest {
	Transform[ ] transforms;
	string algorithm;
	byte[ ] value;
	}
	class Reference {
	string id;
	{Digest digest;}
	}

[0383] 1.21.1.1. Attributes

[0384] In one embodiment, there are four kinds of attributes: IntegerAttribute, StringAttribute, ByteArrayAttribute, and ListAttribute, each having a name and a type.

	abstract class Attribute {
	{string name;}
	string type;
	}
	class IntegerAttribute extends Attribute(type=’int’) {
	int value;
	}
	class StringAttribute extends Attribute(type=’string’) {
	string value;
	}
	class ByteArrayAttribute extends Attribute(type=’bytes’) {
	byte[ ] value;
	}
	Class ListAttribute extends Attribute(type=’list’) {
	Attribute[ ] attributes; // must all be named
	}
	Class ArrayAttribute extends Attribute(type=’array’) {
	Attribute[ ] attributes; // must all be unnamed
	}

[0385] 1.21.1.2. Extensions

[0386] In the illustrative embodiment under discussion, there are two types of extensions: internal extensions, which are carried inside the Octobject, and external extensions, which are carried outside the Octobject.

	abstract class ExtensionData {
	string type;
	}
	abstract class Extension {
	string id;
	}
	class ExternalExtension extends Extension {
	string subject;
	ExtensionData data;
	}
	class InternalExtension extends Extension {
	boolean critical;
	{Digest dataDigest;}
	(ExtensionData data;)
	}

[0387] In some embodiments, it will be important to be able to verify the signature of an object even if a particular type of ExtensionData is not understood by a given implementation. Thus, in one embodiment, a level of indirection with the dataDigest field is added. If the specification of the ExtensionData mandates that the data is part of the signature within the context of a particular object, then the dataDigest field will be present. An implementation that understands this ExtensionData, and is therefore capable of computing its canonical representation, can then verify the digest. If, in such an embodiment, the specification of this ExtensionData mandates that the data is not part of the signature, then the dataDigest field will not be present.

[0388] 1.21.2. Node Objects

	class Node extends Octobject {
	}

[0389] 1.21.3. Link Objects

	class Link extends Octobject {
	string fromId;
	string toId;
	{Control control;}
	}

[0390] 1.21.4. Control Objects

	class Control extends Octobject {
	string protocol;
	string type;
	byte[ ] codeModule;
	}

[0391] 1.21.5. ContentKey Objects

	abstract class Key {
	string id;
	string usage;
	string format;
	byte[ ] data;
	}
	abstract class PairedKey extends Key {
	string pairId;
	}
	class ContentKey extends Octobject {
	Key secretKey;
	}

[0392] In one embodiment, each key has a unique id, a format, a usage (that can be null), and data. The ‘usage’ field, if it is not empty, specifies the purpose for which the key can be used. For normal content keys, this field is empty. In embodiments in which a key distribution scheme such as that described above is used, this field may specify if this is a sharing key or a confidentiality key. The ‘format’ field specifies the format of the ‘data’ field (such as, for example, ‘RAW’ for symmetric keys, or ‘PKCS#8’ for RSA private keys, etc.). The ‘data’ field contains the actual key data, formatted according to the ‘format’ field.

[0393] For keys that are part of a key pair (such as RSA keys), the extra field ‘pairId’ gives a unique identifier for the pair, so that the pair can be referenced from other data structures.

[0394] In one embodiment the data field in the key object is the plaintext value of the actual key (i.e., it is the plaintext value of the key that will be hashed), even though the object's actual representation contains an encrypted copy of the key.

[0395] 1.21.6. Controller Objects

	class Controller extends Octobject {
	Reference controlRef;
	Reference[ ] contentKeyRefs;
	}

8. Virtual Machine

[0396] Preferred embodiments of the DRM engine described herein use a virtual machine (sometimes referred to herein as the “control virtual machine,” the “control VM,” or simply the “VM”) to execute control programs that govern access to content. Illustrative embodiments of such a virtual machine are described below, as are various modifications and design considerations that could be made to this illustrative embodiment. The integration of an illustrative embodiment of the virtual machine (referred to as the “Plankton” virtual machine) with an illustrative embodiment of the DRM engine (referred to as “Octopus”) is also described. It should be appreciated, however, that embodiments of the digital rights management engine, architecture, and other systems and methods described herein can be used with any suitable virtual machine, or, in some embodiments, without a virtual machine at all, and thus it will be appreciated that the details provided below regarding example embodiments of a virtual machine are for purposes of illustration and not limitation.

[0397] In a preferred embodiment, the control VM is a traditional virtual machine, designed to be easy to implement using various programming languages with a very small code footprint. It is based on a simple, stack-oriented instruction set that is designed to be minimalist, without undue concern for execution speed or code density. In situations where compact code is required, data compression techniques can be used to compress the virtual machine's byte code.

[0398] In preferred embodiments, the control virtual machine is designed to be suitable as a target for low or high level programming languages, and supports assembler, C, and FORTH. In addition, it will be appreciated that compilers for other languages, such as Java or custom languages, can be created in a relatively straightforward fashion to compile code into the format (e.g., byte code) used by the virtual machine. In one embodiment the control virtual machine is designed to be hosted within a host environment, not run directly on a processor or in silicon. In preferred embodiments, the natural host environment for the virtual machine is the DRM engine, although it will be appreciated that the virtual machine architecture described herein could alternatively, or in addition, be used in other contexts.

[0399] FIG. 29 illustrates the operating environment of an illustrative implementation of the control virtual machine 2902. As shown in FIG. 29, in one embodiment virtual machine 2902 runs within the context of its host environment 2904, which implements some of the functions needed by the virtual machine as it executes programs 2906. Typically, the control VM runs within the DRM engine 2908, which implements its host environment. As shown in FIG. 29, in a preferred database, the virtual machine 2902 and the DRM engine 2908 have access to a secure database 2910 for presistant storage of state information.

[0400] 1.22. Architecture

[0401] 1.22.1. Execution Model

[0402] In preferred embodiments, the VM runs programs by executing instructions stored in byte code in code modules. Some of these instructions can call functions implemented outside of the program itself by making system calls. System calls can be implemented by the VM itself or delegated to the host environment.

[0403] In one embodiment, the VM executes instructions stored in code modules as a stream of byte codes loaded into memory. The VM maintains a virtual register called the Program Counter (PC), which is incremented as instructions are executed. The VM executes each instruction, in sequence, until an OP_STOP instruction is encountered, an OP_RET instruction is encountered with an empty call stack, or a runtime exception occurs. Jumps are specified either as a relative jump (specified as a byte offset from the current value of PC), or as an absolute address.

[0404] 1.22.2. Memory Model

[0405] In one embodiment, the VM uses a relatively simple memory model, in which memory is separated into data memory and code memory. For example, data memory can be implemented as a single, flat, contiguous memory space, starting at address 0, and can be implemented as an array of bytes allocated within the heap memory of the host application or host environment. In one embodiment, attempts to access memory outside of the allocated space will cause a runtime exception which will cause program execution to terminate.

[0406] Data memory is potentially shared between several code modules concurrently loaded by the virtual machine. The data in the data memory can be accessed by memory-access instructions, which, in one embodiment, can be either 32-bit or 8-bit accesses. 32-bit memory accesses are performed using big-endian byte order. In a preferred embodiment, no assumptions are made with regards to alignment between the virtual machine-visible memory and the host-managed memory (i.e., the host CPU virtual or physical memory).

[0407] In one embodiment, code memory is a flat, contiguous memory space, starting at address 0, and can be implemented as an array of bytes allocated within the heap memory of the host application or host environment.

[0408] The VM may support loading more than one code module. If the VM loads several code modules, in one embodiment all the code modules share the same data memory (although each module's data is preferably loaded at a different address), but each has its own code memory, thus preventing a jump instruction in one code module to cause a jump to code in another code module.

[0409] 1.22.3. Data Stack

[0410] In one embodiment, the VM has the notion of a data stack, which represents 32-bit data cells stored in the data memory. The VM maintains a virtual register called the Stack Pointer (SP). After reset, the SP points to the end of the data memory, and the stack grows downward (when data is pushed on the data stack, the SP register is decremented). The 32-bit data cells on the stack are interpreted either as 32-bit addresses or 32-bit integers, depending on the instruction referencing the stack data. Addresses are unsigned integers. In one embodiment, all other 32-bit integer values on the data stack are interpreted as signed integers unless otherwise specified.

[0411] 1.22.4. Call Stack

[0412] In one embodiment, the VM manages a call stack used for making subroutine calls. In one embodiment, the values pushed on this stack cannot be read or written directly by the memory-access instructions. This stack is used internally by the VM when executing OP_JSR, OP_JSRR, and OP_RET instructions. For a given VM implementation, the size of this return address stack can be fixed to a maximum, which will allow only a certain number of nested calls.

[0413] 1.22.5. Pseudo Registers

[0414] In one embodiment, the VM reserves a small address space at the beginning of data memory to map pseudo-registers. In one embodiment, the addresses of these pseudo-registers are fixed. For example, the following registers could be defined:

Address	Size	Name	Description
0	4	ID	32-bit ID of the currently executing code
			segment. This ID is chosen by the VM
			when a module is loaded. The VM
			changes this register if it changes from
			the code segment of one module to the
			code segment of another module
4	4	DS	32-bit value set to the absolute data
			address at which the data segment of the
			currently executing module has been
			loaded. This value is determined by the
			VM's module loader
8	4	CS	32-bit value set to the absolute code
			address at which the code segment of the
			currently executing module has been
			loaded. This value is determined by the
			VM's module loader.
12	4	UM	32-bit value set to the absolute data
			address of the first byte following the
			region of the data memory space where
			the data segments of code modules have
			been loaded.

[0415] 1.22.6. Memory Map

[0416] The following shows the layout of data memory and code memory in an illustrative embodiment:

[0417] Data Memory

Address Range	Description
Data Memory
0 to 15	Pseudo-registers
16 to 127	Reserved for future VM/System use
128 to 255	Reserved for application use
256 to DS − 1	Unspecified. The VM may load the data
	segments of code modules at any address at or
	above 256. If it chooses an address larger than
	256, the use of the address space between 256
	and DS is left unspecified. This means that the
	virtual machine implementation is free to use it
	any way it sees fit.
DS to UM − 1	Image of the data segments of one or more code
	modules loaded by the virtual machine.
UM to End	Shared address space. The code modules' data
	and the data stack share this space. The data stack
	is located at the end of that space and grows
	down. The end represents the last address of the
	data memory space. The size of the data memory
	space is fixed by the VM based on memory
	requirements contained in the code module and
	implementation requirements.
Code Memory
0 to CS − 1	Unspecified. The virtual machine may load the
	code segments of code modules at any address at
	or above 0. If it chooses an address larger than 0,
	the use of the address space between 0 and CS is
	left unspecified. This means that the virtual
	machine is free to use it in any way it sees fit.
CS to CS + size (code	Image of the code segment of a code module
segment) − 1	loaded by the virtual machine

[0418] 1.22.7. Executing Routines

[0419] Before executing a code routine, in one embodiment the virtual machine implementation resets the data stack pointer to point to the top of the initialized data stack. The initialized data stack contains the routine's input data, and extends to the end of the data memory. The initialized data stack may be used as a way to pass input arguments to a routine. When there is no initialized data stack, the data stack pointer points to the end of the data memory. In one embodiment, the initial call stack is either empty or contains a single terminal return address pointing to an OP_STOP instruction, which will force execution of the routine to end on an OP_STOP instruction in case the routine finished with an OP_RET instruction.

[0420] When execution stops, either because a final OP_RET instruction with an empty call stack has been executed or a final OP_STOP instruction has been executed, any data left on the data stack is considered to be the output of the routine.

[0421] 1.22.8. Runtime Exceptions

[0422] In one embodiment, any of the following conditions is considered to be a runtime exception which causes execution to stop immediately:

• • [0423] An attempt to access data memory outside the current data memory address space.
  • [0424] An attempt to set the PC to, or cause the PC to, reach a code address outside the current code memory address space.
  • [0425] An attempt to execute undefined byte code.
  • [0426] An attempt to execute an OP_DIV instruction with a top-of-stack operand equal to 0.
  • [0427] An attempt to execute an OP_MOD instruction with a top-of-stack operand equal to 0.
  • [0428] An overflow or underflow of the Call Stack.

[0429] 1.23. Instruction Set

[0430] In one embodiment, the control VM uses a relatively simple instruction set. Though limited, the number of instructions is sufficient to express programs of arbitrary complexity. Instructions and their operands are represented by a stream of byte codes. In one embodiment, the instruction set is stack-based, and except for the OP_PUSH instruction, none of the instructions have direct operands. Operands are read from the data stack, and results pushed on the data stack. In one embodiment, the VM is a 32-bit VM: all the instructions operate on 32-bit stack operands, representing either memory addresses or signed integers. Signed integers are represented with 2 s complement binary encoding. An illustrative embodiment of an instruction set for use with the control VM is shown in the following table. In the table, the stack operands for instructions with two operands are listed as “A,B” where the operand on the top of the stack is listed last (i.e., “B”). Unless otherwise specified, the term “push,” as used in the following description of one illustrative embodiment, refers to pushing a 32-bit value onto the top of the data stack.

		Byte
OP CODE	Name	Code	Operands	Description
OP_NOP	No	0		Do Nothing
	Operation
OP_PUSH	Push	1	N (direct)	Push a 32-bit constant
	Constant
OP_DROP	Drop	2		Remove the top cell of the data
				stack
OP_DUP	Duplicate	3		Duplicate the top cell of the
				data stack
OP_SWAP	Swap	4		Swap top two stack cells
OP_ADD	Add	5	A, B	Push the sum of A and B (A + B)
OP_MUL	Multiply	6	A, B	Push the product of A and B
				(A * B)
OP_SUB	Subtract	7	A, B	Push the difference between A
				and B (A − B)
OP_DIV	Divide	8	A, B	Push the division of A by B
				(A/B)
OP_MOD	Modulo	9	A, B	Push A modulo B (A % B)
OP_NEG	Negate	10	A	Push the 2's complement
				negation of A (−A)
OP_CMP	Compare	11	A, B	Push −1 if A less than B, 0 if A
				equals B, and 1 if A greater
				than B
OP_AND	And	12	A, B	Push bit-wise AND of A and B
				(A & B)
OP_OR	Or	13	A, B	Push the bit-wise OR of A and
				B (A | B)
OP_XOR	Exclusive	14	A, B	Push the bit-wise eXclusive OR
	Or			of A and B (A B)
OP_NOT	Logical	15	A	Push the logical negation of A
	Negate			(1 if A is 0, and 0 if A is not 0)
OP_SHL	Shift Left	16	A, B	Push A logically shifted left by
				B bits (A << B)
OP_SHR	Shift Right	17	A, B	Push A logically shifted right
				by B bits (A >> B)
OP_JMP	Jump	18	A	Jump to A
OP_JSR	Jump to	19	A	Jump to subroutine at absolute
	Subroutine			address A. The current value of
				PC is pushed on the call stack.
OP_JSRR	Jump to	20	A	Jump to subroutine at PC + A.
	Subroutine			The current value of PC is
	(Relative)			pushed on the call stack.
OP_RET	Return from	21		Return from subroutine to the
	Subroutine			return address popped from the
				call stack.
OP_BRA	Branch	22	A	Jump to PC + A
	Always
OP_BRP	Branch if	23	A, B	Jump to PC + B if A > 0
	Positive
OP_BRN	Branch if	24	A, B	Jump to PC + B if A < 0
	Negative
OP_BRZ	Branch if	25	A, B	Jump to PC + B if A is 0
	Zero
OP_PEEK	Peek	26	A	Push the 32-bit value at address A
OP_POKE	Poke	27	A, B	Store the 32-bit value A at
				address B
OP_PEEKB	Peek Byte	28	A	Read the 8-bit value at address
				A, 0-extend it to 32-bits and
				push it on the data stack
OP_POKEB	Poke Byte	29	A, B	Store the least significant 8 bits
				of value A at address B
OP_PUSHSP	Push Stack	30		Push the value of SP
	Pointer
OP_POPSP	Pop Stack	31	A	Set the value of SP to A
	Pointer
OP_CALL	System Call	32	A	Perform System Call with
				index A
OP_STOP	Stop	255		Terminate Execution

[0431] 1.24. Code Modules

[0432] In a preferred embodiment, code modules are stored in an atom-based format, similar or identical to that used for the MPEG-4 file format, in which atoms contain a 32-bit size (e.g., represented by 4 bytes in big-endian byte order), followed by a 4-byte type (e.g., bytes that correspond to ASCII values of letters of the alphabet), followed by a payload (e.g., 8 bytes).

[0433] FIG. 30 shows the format of an illustrative code module 3000. Referring to FIG. 30, pkCM atom 3002 is the top-level code module atom. It contains a sequence of sub-atoms. In one embodiment, pkCM atom 3002 contains one pkDS atom 3004, one pkCS atom 3006, one pkEX atom 3008, and possibly one pkRQ atom 3010. The pkCM atom 3002 may also contain any number of other atoms that, in one embodiment, are ignored if present. In one embodiment, the order of the sub-atoms is not specified, so implementations should not assume a specific order.

[0434] 1.24.1. pkDS Atom

[0435] As shown in FIG. 30, pkDS atom 3004 contains a memory image 3005 of a data segment that can be loaded into data memory. As shown in FIG. 31A, in one embodiment memory image 3005 is represented by a sequence of bytes 3112, consisting of one header byte 3114 followed by zero or more data bytes 3116. Header byte 3114 encodes a version number that identifies the format of the bytes that follow 3116.

[0436] In one embodiment, only one version number is defined (i.e., DataSegmentFormatVersion=0), and in this format the data bytes of the memory image represent a raw image to be loaded into memory. The virtual machine loader only loads the data bytes 3116 of the memory image 3105, not including the header byte 3114. In one embodiment, the virtual machine loader is operable to refuse to load an image in any other format.

[0437] 1.24.2. pkCS Atom

[0438] As shown in FIG. 30, pkCS atom 3006 contains a memory image 3007 of a code segment that can be loaded into code memory. As shown in FIG. 31B, in one embodiment memory image 3007 is represented by a sequence of bytes 3120 consisting of one header byte 3122 followed by zero or more data bytes 3124. Header byte 3122 encodes a version number that identifies the format of the bytes that follow 3124.

[0439] In one embodiment, only one version number is defined (i.e., CodeSegmentFormatVersion=0), and, as shown in FIG. 31C, in this version the byte following header byte 3122 contains another header byte 3130 containing a version number that identifies the byte code encoding of the following bytes 3132. In the example shown in FIG. 31C, header byte 3130 identifies ByteCodeVersion=0, which specifies that data bytes 3132 contain a raw byte sequence with byte code values such as those defined in the example instruction set that is set forth above. In a preferred embodiment, the virtual machine loader only loads the byte code portion 3132 of the data bytes, not the two header bytes 3122, 3130.

[0440] 1.24.3. pkEX Atom

[0441] Referring once again to FIG. 30, the pkEX atom 3008 contains a list of export entries. In the example shown in FIG. 30, the first four bytes 3009 of pkEX atom 3008 encode a 32-bit unsigned integer in big-endian byte order equal to the number of entries that follow. As shown in FIG. 31D, each following export entry 3160 consists of a name, encoded as one byte 3162 containing the name size, S, followed by S bytes 3164 containing the ASCII characters of the name, including a terminating zero 3166, followed by a 32-bit unsigned integer 3168 in big-endian byte order representing the byte offset of the named entry point, measured from the start of the byte code data stored in the 3 ICS atom. FIG. 31E shows an example of an export table entry 3170 for the entry point MAIN at offset 64, in which the first byte 3172 indicates that the size of the name (i.e., “MAIN”), plus the terminating zero, is five bytes, and in which the last four bytes 3174 indicate that the byte offset is 64.

[0442] 1.24.4. pkRQ Atom

[0443] As shown in FIG. 30, pkRQ atom 3010 contains requirements that need to be met by the virtual machine implementation in order to execute the code in the code module. In one embodiment, this atom is optional, and if it is not present, the virtual machine uses default implementation settings, such as may be defined by an implementation profile.

[0444] In one embodiment, the pkRQ atom consists of an array of 32-bit unsigned integer values, one for each field:

Field Name	Description
vmVersion	Version ID of the VM Spec
minDataMemorySize	Minimum size in bytes of the data memory
	available to the code. This includes the data
	memory used to load the image of the Data
	Segment, as well as the data memory used by
	the Data Stack. In one embodiment, the VM
	must refuse to load the module if it cannot
	satisfy this requirement.
minCallStackDepth	Minimum number of nested subroutine calls
	(OP_JSR and OP_JSRR) that must be
	supported by the VM. In one embodiment,
	the VM must refuse to load the module if it
	cannot satisfy this requirement.
Flags	Set of bit-flags to signal required features of
	the VM.
	In one embodiment, a VM implementation
	must refuse to load a code module that has
	any unknown flag set. For example, if there
	are no flags defined, in one embodiment a
	VM implementation must check that this flag
	is set to 0.

[0445] 1.24.5. Module Loader

[0446] The virtual machine is responsible for loading code modules. When a code module is loaded, the Data Segment memory image encoded in the pkDS atom is loaded at a memory address in the Data Memory. That address is chosen by the VM loader, and is stored in the DS pseudo-register when the code executes.

[0447] The Code Segment memory image encoded in the pkCS atom is loaded at a memory address in the Code Memory. That address is chosen by the VM loader, and is stored in the CS pseudo-register when the code executes.

[0448] When a code module is loaded, the special routine named “Global.OnLoad” is executed if this routine is found in the entries of the Export table. This routine takes no argument on the stack, and returns an integer status upon return, 0 signifying success, and a negative error code signifying an error condition.

[0449] When a code module is unloaded (or when the virtual machine that has loaded the module is disposed of), the special routine named “Global.OnUnload” is executed if it is found in the Export table. This routine takes no argument on the stack, and returns an integer status upon return, 0 signifying success, and a negative error code signifying an error condition.

[0450] 1.25. System Calls

[0451] The virtual machine's programs can call functions implemented outside of their code module's Code Segment. This is done through the use of the OP_CALL instruction, which takes an integer stack operand specifying the System Call Number to call. Depending on the System Call, the implementation can be a byte code routine in a different code module (for instance, a library of utility functions), executed directly by the VM in the VM's native implementation format, or delegated to an external software module, such as the VM's host environment.

[0452] In one embodiment, if an OP_CALL instruction is executed with an operand that contains a number that does not correspond to any System Call, the VM behaves as if the SYS_NOP system call was called.

[0453] 1.25.1. System Call Numbers Allocation

[0454] In the illustrative embodiment under discussion, System Call Numbers 0 to 1023 are reserved for fixed System Calls (these System Calls will have the same number on all VM implementations). System Call Numbers 1024 to 16383 are available for the VM to assign dynamically (for example, the System Call Numbers returned by System.FindSystemCallByName can be allocated dynamically by the VM, and do not have to be the same numbers on all VM implementations).

[0455] In one example embodiment, the following fixed System Call Numbers are specified:

Mnemonic	Number	System Call
SYS_NOP	0	System.NoOperation
SYS_DEBUG_PRINT	1	System.DebugPrint
SYS_FIND_SYSTEM_CALL_BY_NAME	2	System.FindSystemCallByName
SYS_SYSTEM_HOST_GET_OBJECT	3	System.Host.GetObject
SYS_SYSTEM_HOST_SET_OBJECT	4	System.Host.SetObject

[0456] 1.25.2. Standard System Calls

[0457] In one embodiment, a few standard system calls are supported that are useful for writing control programs. These calls include the fixed-number system calls listed in the table above, as well as system calls that have dynamically determined numbers (i.e., their system call number is retrieved by calling the System.FindSystemCallByName system call with their name passed as the argument).

[0458] In one embodiment, the system calls specified in this section that can return a negative error code may return error codes with any negative value. Section 8.4.4 defines specific, illustrative values. In one embodiment, if negative error code values are returned that are not predefined, they are interpreted as if they were the generic error code value FAILURE.

[0459] System.NoOperation. This call takes no inputs and returns no outputs, and simply returns without doing anything. It is used primarily for testing the VM.

[0460] System.DebugPrint. This call takes as its input, from the top of the stack, the address of a memory location containing a null-terminated string, and returns no output. A call to this function causes the string of text to be printed to a debug output, which can be useful in debugging. If the VM implementation does not include a facility to output debug text (such as might be the case in a non-development environment), the VM may ignore the call and treat it as if System.NoOperation had been called.

[0461] System.FindSystemCallByName. This call finds the number of a system call given its name. The call takes as its input (from the top of the stack) the address of a null-terminated ASCII string containing the name of the system call for which to look, and returns (to the top of the stack) the system call number if a system call with the specified name is implemented, an ERROR_NO_SUCH_ITEM if the system call is not implemented, and a negative error code if an error occurs.

[0462] System.Host.GetLocalTime. This call takes no inputs, and returns, to the top of the stack, the current value of the local time of the host, which, in one embodiment, is expressed as a 32-bit signed integer equal to the number of minutes elapsed since Jan. 1, 1970 00:00:00, or a negative error code.

[0463] System.Host.GetLocalTimeOffset. This call takes no inputs, and returns, to the top of the stack, the current time offset (from UTC time) of the host, which, in one embodiment, is expressed as a 32-bit signed integer number equal to the number of minutes difference between local time and UTC time (i.e. LocalTime—UTC).

[0464] System.Host.GetTrustedTime. This call takes no inputs, and returns, to the top of the stack, the trusted time and the value of one or more flags. In one embodiment, the trusted time is the current value of the trusted time clock (if the system includes such a trusted clock), or a negative error code if the trusted time is not available. In one embodiment, the value of trusted time is expressed as a 32-bit signed integer equal to the number of minutes elapsed since Jan. 1, 1970 00:00:00 UTC, or a negative error code. In one embodiment the flags are the bit-set of flags that further define the current state of the trusted clock. In one embodiment, if an error has occurred (e.g., the value of TrustedTime is a negative error code) the value returned for the flags is 0.

[0465] In one embodiment, the following flag is defined:

Bit index (0 is
LSB)	Name	Description
0	TIME_IS_ESTIMATE	The value of TrustedTime is
		known to not be at its most
		accurate value, and therefore
		should be considered
		an estimate.

[0466] This system call is relevant on systems that implement a trusted clock that can be synchronized with a trusted time source and maintain a monotonic time counter. The value of the trusted time is not guaranteed to always be accurate, but in one embodiment the following properties are required to be true:

• • [0467] The value of the trusted time is expressed as a UTC time value (the trusted time is not in the local time zone, as the current locality usually cannot be securely determined).
  • [0468] The trusted time never goes back.
  • [0469] The trusted clock does not advance faster than realtime.

[0470] Therefore, in this example embodiment, the value of TrustedTime is between the value of the last synchronized time (synchronized with a trusted time source) and the current real time. If the system is able to determine that its trusted clock has been operating and updating continuously and normally without interruption since the last synchronization with a trusted time source, it can determine that the value of TrustedTime is not an estimate, but an accurate value, and set the TIME_IS_ESTIMATE flag to 0.

[0471] In one embodiment, if the trusted clock detects that a hardware or software failure condition has occurred, and it is unable to return even an estimate of the trusted time, an error code is returned, and the value of the returned flags is set to 0.

[0472] System.Host.GetObject: This system call is a generic interface that allows a program to access objects provided by the virtual machine's host. The System.Host.GetObject call takes the following inputs (listed from the top of the stack downwards): Parent, Name, ReturnBuffer, and ReturnBuffer Size. Where “Parent” is the 32-bit handle of the parent container; “Name” is the address of a null-terminated string containing the path to the requested object, relative to the parent container; “ReturnBuffer” is the address of a memory buffer where the value of the object is to be stored; and “ReturnBufferSize” is a 32-bit integer indicating the size in bytes of the memory buffer in which the value of the object is to be stored.

[0473] The System.Host.GetObject call produces the following outputs (listed from the top of the stack downwards): TypeID, Size. Where “TypeId” is the object type id, or a negative error code if the call failed. If the requested object does not exist, the error returned is ERROR_NO_SUCH_ITEM. If the buffer supplied for the return value is too small, the error returned is ERROR_INSUFFICIENT_SPACE. If the part of the object tree that is being accessed is access-controlled, and the calling program does not have the permission to access the object, ERROR_PERMISSION_DENIED is returned. Other error codes may be returned. “Size” is a 32-bit integer indicating the size in bytes of the data returned in the buffer supplied by the caller, or the size required if the caller provided a buffer that was too small.

[0474] In one embodiment, there are four types of host objects: strings, integers, byte arrays, and containers.

Object Type	Type Id Name	Type Id Value
Container	OBJECT_TYPE_CONTAINER	0
Integer	OBJECT_TYPE_INTEGER	1
String	OBJECT_TYPE_STRING	2
Byte Array	OBJECT_TYPE_BYTE_ARRAY	3

[0475] In one embodiment, the value of a byte array object is an array of 8-bit bytes, the value of a string object is a null-terminated character string incoded in UTF-8, and the value of an integer object is a 32-bit signed integer value. Containers are generic containers that contain a sequence of any number of objects of any combination of types. Objects contained in a container are called the children of that container. The value of a container is a 32-bit container handle that is unique within a given VM instance. In one embodiment, the root container ‘/’ has the fixed handle value 0.

[0476] In one embodiment, the namespace for host objects is hierarchical, where the name of a container's child object is constructed by appending the name of the child to the name of the parent container, separated by a ‘/’ character. String and integer objects do not have children. For example, if a container is named ‘/Node/Attributes’, and has a string child named ‘Type’, then ‘/Node/Attributes/Type’ refers to the child string.

[0477] The root of the namespace is ‘/’. All absolute names start with a ‘/’. Names that do not start with a ‘/’ are relative names. Relative names are relative to a parent container. For example, the name ‘Attributes/Type’, relative to parent ‘/Node’, is the object with the absolute name ‘/Node/Attributes/Type’.

[0478] In one embodiment, container objects can also have real and virtual child objects that be accessed by using virtual names. Virtual names are names that are not attached to host objects, but a convention to identify either unnamed child objects, child objects with a different name, or virtual child objects (child objects that are not real children of the container, but created dynamically when requested).

[0479] In one embodiment, for objects, the following virtual names are defined as virtual child object names:

	Virtual Name	Description
	@Name	Virtual string object: the name of the object.
		If the object is unnamed, the value is an empty
		string. Note that unnamed objects are only
		accessible through the @<n> virtual name of a
		container object (see below)
	@Size	Virtual integer object. The integer value is equal
		to the size in bytes required to store this object.
		For integers, this value is 4; for strings, it is the
		number of bytes needed to store the UTF-8
		string plus a null byte terminator. For byte
		arrays, this is the number of bytes in the array.
	@Type	Virtual integer object. The integer value is equal
		to the object's Type Id.

[0480] For containers, the following virtual names are defined as virtual child object names in one embodiment:

	Virtual Name	Description
Virtual	@<n>	Virtual object: the <n>th object in a container.
Index		The first object in a container has index 0. <n>
		is expressed as a decimal number.
		Example: if ‘Attributes’ is a container that
		contains 5 child objects, ‘Attributes/@4’ is the
		5th child of the container.
Virtual	@Size	Virtual integer object. The integer value is
Size		equal to the number of objects in the container.

[0481] The following table shows an example of a hierarchy of Host Objects:

		Chil-
Name	Value	dren
Node	1	Name	Value	Children
		Type	“Device”
		Name	Value	Children
		Attri-	2	Name	Value	Children
		butes		Color	“Red”
				Name	Value	Children
				Size	78
				Name	Value	Children
				Domain	“TopLevel”

[0482] In this example, calling System.Host.GetObject(parent=0, name=“Node”) returns a type ID of 0 (i.e., container), and causes the handle value of 1 to be written in the buffer supplied by the caller. The size of the value is 4 bytes.

[0483] Calling System.Host.GetObject(parent=0, name=“Node/Attributes/Domain”) returns a type ID of 2 (i.e., string), and causes the string “TopLevel” to be written in the buffer supplied by the caller. The size of the value is 9 bytes.

[0484] Calling System.Host.GetObject(parent=1, name=“Attributes/@ 1”) returns a type ID of 1 (i.e., integer), and causes the integer 78 to be written in the buffer supplied by the called. The size of the value is 4 bytes.

[0485] Calling System.Host.GetObject(parent=00, name=“DoesNotExist”) returns the error code ERROR_NO_SUCH_ITEM.

[0486] System.Host.SetObject. This system call is a generic interface that allows a program to create, write, and destroy objects provided by the virtual machine's host. The description of the object names and types is the same as for the System.Host.GetObject call described above. Not all host objects support being written to or destroyed, and not all containers support having child objects created. When a SetObject call is made for an object that does not support the operation, ERROR_PERMISSION_DENIED is returned.

[0487] The System.Host.SetObject system call takes as input the following parameters, listed from the top of the stack downwards:

Top of stack
	Parent
	Name
	ObjectAddress
	ObjectType
	ObjectSize
	. . .

[0488] Parent: 32-bit handle of the parent container.

[0489] Name: address of a null-terminated string containing the path to the object, relative to the parent container.

[0490] ObjectAddress: address of a memory buffer where the value of the object is stored. If the address is 0, the call is interpreted as a request to destroy the object. The data at the address depends on the type of the object.

[0491] ObjectType: the type ID of the object.

[0492] ObjectSize: 32-bit integer indicating size in bytes of the memory buffer where the value of the object is stored. In the illustrative embodiment under discussion, the size is set to 4 for integer objects, and to the size of the memory buffer, including the null terminator, for string objects. For byte array objects, the size is the number of bytes in the array.

[0493] The System.Host.SetObject system call returns a ResultCode to the top of the stack as an output. The ResultCode is 0 if the call succeeded, and a negative error code if the call failed. If the call is a request to destroy an object and the requested object does not exist, or the call is a request to create or write an object and the object's parent does not exist, the error code returned is ERROR_NO_SUCH_ITEM. If the part of the object tree that is being accessed is access-controlled, and the calling program does not have the permission to access the object, ERROR_PERMISSION_DENIED is returned. Other error codes may also be returned.

[0494] There is a special case when the object refers to a container and the ObjectAddress is not 0. In this case the ObjectSize parameter is set to 0 and the value of ObjectAddress is ignored. If the container already exists, nothing is done, and a SUCCESS ResultCode is returned. If the container does not exist, and the parent of the container is writeable, an empty container is created.

[0495] Octopus.Links.IsNodeReachable. This system call is used by control programs to check whether a given node is reachable from the node associated with the entity hosting this instance of the virtual machine. The call takes as its input a NodeId from the top of the stack, where the NodeId is a null-terminated string containing the ID of the target node to be tested for reachability. As output, the call returns a ResultCode and a StatusBlockPointer to the top of the stack. The ResultCode is an integer value that is 0 if the node is reachable, or a negative error code if it is not. The StatusBlockPointer is the address of a standard ExtendedStatusBlock, or 0 if no status block is returned.

[0496] System.Host.SpawnVm. This system call is used by control programs to request that a new instance of a virtual machine be created, and a new code module loaded. In one embodiment, the host of the newly created virtual machine exposes the same host objects as the ones exposed to the caller, except the host object “/Octopus/Runtime/Parent/Id” is set to the identity of the caller. In one embodiment, this host object is a container. The children of this container are objects of type string, each with a value representing a name. In one embodiment, the semantics and specific details of those names are specified by the specification of the virtual mchine's host.

[0497] In one embodiment, when the virtual machine that is running the code for the caller terminates, any spawned virtual machine that has not been explicitly released by calling System.Host.ReleaseVm is automatically released by the system as if System.Host.ReleaseVm had been called.

[0498] The System.Host.SpawnVm call takes as its input a ModuleId from the top of the stack. The ModuleId identifies the code module to be loaded into the new virtual machine instance. In one embodiment, the specification of the virtual machine's host describes the mechanism by which the actual code module corresponding to this module ID is to be located.

[0499] The System.Host.SpawnVm call returns a ResultCode and a VmHandle to the top of the stack. The ResultCode is an integer value that is 0 if the call was successful, and a negative error code if it failed. The VmHandle is an integer value identifying the instance of the virtual machine that has been created. If the call fails, this handle is set to 0. In one embodiment, this handle is only guaranteed to be unique within the virtual machine in which this call is made.

[0500] System.Host.CallVm. This system call is used by control programs to call routines that are implemented in code modules loaded in virtual machine instances created using the System.Host.SpawnVm system call. This system call takes the following input from the top of the stack:

Top of stack:
	VmHandle
	EntryPoint
	ParameterBlockAddress
	ParameterBlockSize
	ReturnBufferAddress
	ReturnBufferSize
	. . .

[0501] VmHandle: an integer value representing the handle of a virtual machine that was created by calling System.Host.SpawnVm.

[0502] EntryPoint: the address of a null-terminated string that specifies the name of the entry point to call. This name needs to match one of the entry points in the Export Table of the code module that was loaded into the virtual machine instance that corresponds to the VmHandle parameter.

[0503] ParameterBlockAddress: the address of a memory block that contains data to be passed to the callee. If no parameters are passed to the callee, this address is set to 0.

[0504] ParameterBlockSize: the size in bytes of the memory block at address ParameterBlockAddress, or 0 if ParameterBlockAddress is 0.

[0505] ReturnBufferAddress: the address of a memory buffer where the caller can receive data from the callee. If the caller does not expect any data back from the callee, this address is set to 0.

[0506] ReturnBufferSize: the size in bytes of the memory buffer at address ReturnBufferAddress, or 0 if ReturnBufferAddress is 0.

[0507] The System.Host.CallVm call returns the following output to the top of the stack:

Top of Stack:
	SystemResultCode
	CalleeResultCode
	ReturnBlockSize
	. . .

[0508] SystemResultCode: an integer value that is 0 if the call was successful or a negative error code if it failed. This value is determined by the system, not by the callee. Success only indicates that the system was able to successfully find the routine to call, execute the routine, and get the return value from the routine. The return value from the routine itself is returned in the CalleeResultCode value.

[0509] CalleeResultCode: an integer value that is returned by the callee.

[0510] ReturnBlockSize: the size in bytes of the data returned in the buffer supplied by the caller, or the size required if the caller provided a buffer that was too small. If no data was returned by the callee, the value is 0.

[0511] In the illustrative embodiment under discussion, the called routine complies with the following interface conventions: When the routine is called, the top of the stack contains the value ParameterBlockSize, supplied by the caller, indicating the size of the parameter block, followed by ParameterBlockSize bytes of data. If the size is not a multiple of 4, the data on the stack will be padded with zeros to ensure that the stack pointer remains a multiple of 4. Upon return, the called routine provides the following return values on the stack:

Top of stack:
	ResultCode
	ReturnBlockAddress
	ReturnBlockSize
	. . .

[0512] ReturnBlockAddress: the address of a memory block that contains data to be returned to the caller. If no data is returned, this address is set to 0.

[0513] ReturnBlockSize: size in bytes of the memory block at address ReturnBlockAddress, or 0 if ReturnBlockAddress is 0.

[0514] System.Host.ReleaseVm. This system call is used by control programs to release a virtual machine that was spawned by a previous call to System.Host.SpawnVm. Any virtual machines spawned by the released virtual machine are released, and so on, recursively. The System.Host.ReleaseVm call takes as its input a VmHandle from the top of the stack, the VmHandle representing the handle of a virtual machine that was created by calling System.Host.SpawnVm. The System.Host.ReleaseVm call returns a ResultCode to the top of the stack as an output. The ResultCode is an integer value that is 0 if the call was successful or a negative error code if it failed.

[0515] 1.25.3. Standard Data Structures

[0516] The following are standard data structures used by some of the standard system calls.

[0517] 1.25.3.1. Standard Parameters

ParameterBlock:
	Name	Type
	Name	NameBlock
	Value	ValueBlock

[0518] Name: name of the parameter.

[0519] Value: value of the parameter

ExtendedParameterBlock:
	Name	Type
	Flags	32-bit bit field
	Parameter	ParameterBlock

[0520] Flags: vector of boolean flags.

[0521] Parameter: parameter block containing a name and a value.

NameBlock:
	Name	Type
	Size	32-bit integer
	Characters	Array of 8-bit characters

[0522] Size: 32-bit unsigned integer equal to the size in bytes of the “characters” field that follows. If this value is 0, the characters field is left empty (i.e., nothing follows).

[0523] Characters: Null-terminated UTF-8 string.

ValueBlock:
	Name	Type
	Type	32-bit integer
	Size	32-bit integer
	Data	Array of 8-bit bytes

[0524] Type: 32-bit type identifier. In one embodiment, the following types are defined:

Identifier	Type Name	Description
0	Integer	32-bit integer value,
		encoded as four 8-bit bytes
		in big-endian byte order. In
		one embodiment the value
		is considered signed unless
		otherwise specified.
1	Real	32-bit floating point value,
		encoded as IEEE-754 in
		big-endian byte order
2	String	Null-terminated UTF-8
		string
3	Date	32-bit unsigned integer
		value, representing the
		number of minutes elapsed
		since January 1, 1970
		00:00:00. In one
		embodiment, unless
		otherwise specified, the
		value is considered to be a
		UTC date, the most
		significant bit of which
		must be 0.
4	Parameter	ParameterBlock structure
5	ExtendedParameter	ExtendedParameterBlock
		structure
6	Resource	The value is a resource. The
		resource here is referenced
		by ID: the Data field of the
		value is a null-terminated
		ASCII string containing the
		ID of the resource that
		should be de-referenced to
		produce the actual data.
7	ValueList	An array of values (encoded
		as a ValueListBlock)
8	ByteArray	The value is an array of 8-
		bit bytes

[0525] Size: 32-bit unsigned integer equal to the size in bytes of the “data” field that follows. If this value is 0, the data field is left empty (i.e., nothing follows the size field in the ValueBlock).

[0526] Data: array of 8-bit bytes representing a value. The actual bytes depend on the data encoding specified by the type field.

ValueListBlock:
	Name	Type
	ValueCount	32-bit integer
	Value0	ValueBlock
	Value1	ValueBlock
	. . .	. . .

[0527] ValueCount: 32-bit unsigned integer equal to the number of ValueBlock structures that follow. If this value is 0, no ValueBlocks follow.

[0528] Value0, Value1, . . . : sequence of zero or more ValueBlock structures.

[0529] 1.25.3.2. Standard ExtendedStatus

[0530] The standard ExtendedStatusBlock is a data structure typically used to convey extended information as a return status from a call to a routine or a system call. It is a generic data structure that can be used in a variety of contexts, with a range of different possible values for its fields. In one embodiment, an ExtendedStatusBlock is defined as follows:

[0531] ExtendedStatusBlock:

	Name	Type
	GlobalFlags	32-bit bit field
	Category	32-bit integer
	SubCategory	32-bit integer
	LocalFlags	32-bit bit field
	CacheDuration	CacheDurationBlock
	Parameters	ValueListBlock

[0532] GlobalFlags: boolean flags whose semantics are the same regardless of the category field. The position and meaning of the flags are defined by profiles that use standard ExtendedStatusBlock data structures.

[0533] Category: Unique integer identifier of a category to which this status belongs. The category identifier values are defined by profiles that use standard ExtendedStatusBlock data structures.

[0534] SubCategory: Integer identifier (unique within the category) of a sub-category that further classifies the type of status described by this block.

[0535] LocalFlags: Boolean flags whose semantics are local to the category and subcategory of this status block. The position and meaning of the flags are defined by profiles that define and use the semantics of the category.

[0536] CacheDuration: Indicates the duration for which this status can be cached (i.e remains valid). See the definition of the CacheDurationBlock type, below, for how the actual value of the duration is defined.

[0537] Parameters: List of zero or more ValueBlocks. Each ValueBlock contains a parameter encoded as a value of type Parameter or ExtendedParameter. Each parameter binds a name to a typed value, and is used to encode flexible variable data that describes the status block in more detail than just the category, sub-category, cache duration, and flags.

[0538] CacheDurationBlock:

	Name	Type
	Type	32-bit integer
	Value	32-bit integer

[0539] Type: Integer identifier for the type of the value. In one embodiment, the following types are defined:

Type	Description
0	The value is a 32-bit unsigned integer that
	represents the number of seconds from the
	current time. A value of 0 means that the
	status cannot be cached at all, and therefore
	can only be used once. The special value
	0xFFFFFFFF is interpreted as an infinite
	duration (i.e., the status can be cached
	indefinitely).
1	The value is a 32-bit unsigned integer that
	represents an absolute local time, expressed
	as the number of minutes elapsed since
	January 1, 1970 00:00:00. In one
	embodiment, the most significant bit must be
	0.

[0540] Value: 32-bit integer, the meaning of which depends on the Type field.

[0541] 1.25.4. Standard Result Codes

[0542] Standard result codes are used in various APIs. Other result codes may be defined for use in more specific APIs.

Value	Name	Description
0	SUCCESS	Success
−1	FAILURE	Unspecified failure
−2	ERROR_INTERNAL	An internal
		(implementation) error
		has occurred
−3	ERROR_INVALID_PARAMETER	A parameter has an
		invalid value
−4	ERROR_OUT_OF_MEMORY	Not enough memory
		available to complete
		successfully
−5	ERROR_OUT_OF_RESOURCES	Not enough resources
		available to complete
		successfully
−6	ERROR_NO_SUCH_ITEM	The requested item
		does not exist or was
		not found
−7	ERROR_INSUFFICIENT_SPACE	Not enough memory
		supplied by the caller
		(typically used when a
		return buffer is too
		small)
−8	ERROR_PERMISSION_DENIED	Permission to perform
		the call is denied to the
		caller.
−9	ERROR_RUNTIME_EXCEPTION	An error has occurred
		during execution of
		byte code
−10	ERROR_INVALID_FORMAT	Error caused by data
		with an invalid format
		(for example, invalid
		data in a code module)

[0543] 1.26. Assembler Syntax

[0544] This section describes an example syntax for use in compiling programs into the bytecode format described elsewhere herein. It should be appreciated that this is just one example of one possible syntax, and that any suitable syntax could be used. As previously indicated, it should also be understood that the bytecode format presented herein is also just an example, and the systems and methods described herein could be used with any other suitable byte code format or other code format.

[0545] An assembler reads source files containing code, data, and processing instructions, and produces binary code modules that can be loaded by a control virtual machine. In one illustrative embodiment, the assembler processes a source file sequentially, line by line. Lines can be zero or more characters, followed by a newline. Each line can be one of: an empty line (whitespace only), a segment directive, a data directive, an assembler directive, a code instruction, a label, or an export directive. In addition, each line can end with a comment, which starts with a ‘;’ character and continues until the end of the line.

[0546] Data and instructions read from the source files have an implicit destination segment (i.e., where they end up when loaded by the VM). At any point during the parsing process, the assembler will have a “current” segment which is the implicit destination segment for data and instructions. The current segment can be changed using segment directives.

[0547] 1.26.1. Segment Directives

[0548] Segment directives change the current segment of the parser. In one embodiment, the supported segment directives are .code and data. The .code segment holds the byte code instructions, and the data segment holds global variables.

[0549] 1.26.2. Data Directives

[0550] Data directives specify data (e.g., integers and strings) that will be loaded in the virtual machine's data segment. In one embodiment, the supported data directives are:

• • [0551] .string “<some chars>”—Specifies a string of characters. In one embodiment, the assembler adds an octet with value 0 at the end of the string.
  • [0552] .byte <value>—Specifies an 8-bit value. <value> can be expressed as a decimal number, or a hexadecimal number (prefixed by 0x).
  • [0553] .long <value>—Specifies a 32-bit value. <value> can be expressed as a decimal number, or a hexadecimal number (prefixed by 0x).

[0554] 1.26.3. Assembler Directives

[0555] In one embodiment, the supported assembler directives are .equ <symbol>, <value>, which sets the symbol <symbol> to be equal to the value <value>. Symbols are typically used as operands or code instructions.

[0556] 1.26.4. Labels

[0557] Labels are symbols that point to locations within segments. Labels pointing to instructions in the code segment are typically used for jump/branch instructions. Labels pointing to data in the data segment are typically used to refer to variables. In one embodiment, the syntax for a label is: <LABEL>:

[0558] Note that there is nothing after the “:”, except an optional comment. A label points to the location of the next data or instruction. In one embodiment, it is ok to have more than one label pointing to the same address.

[0559] 1.26.5. Export Directives

[0560] Export directives are used to create entries in the “export” section of the code module produced by the assembler. Each entry in the export section is a (name, address) pair. In the illustrative embodiment under discussion, only addresses within the code segment can be specified in the export section.

[0561] The syntax of the export directive is: export <label>, which will export the address pointed to by <label>, with the name “<label>”.

[0562] 1.26.6. Code Instructions

[0563] When compiling data destined for the code segment, the assembler reads instructions that map directly, or indirectly, into byte codes. In the example instruction set shown above, most virtual machine byte codes have no direct operands, and appear with a simple mnemonic on a single line. To make the assembler syntax more readable, some instructions accept pseudo-operands, which look as if they were byte code operands, but are not really; in this case, the assembler generates one or more byte code instructions to produce the same effect as if the instruction did have a direct operand. For example, the branch instructions use pseudo-operands.

[0564] 1.26.6.1. Branch Operands

[0565] Branch instructions can be specified verbatim (without any operand), or with an optional operand that will be converted by the assembler into a corresponding byte code sequence. The optional operand is an integer constant or a symbol. When the operand is a symbol, the assembler computes the correct integer relative offset so that the branch ends up at the address corresponding to the symbol.

[0566] 1.26.6.2. Push Operands

[0567] In one embodiment, the PUSH instruction always takes one operand. The operand can be one of an integer constant, a symbol, or the prefix “@” directly followed by a label name. When the operand is a symbol, the value that is pushed is the direct value of that symbol, whether the symbol is a label or an .equ symbol (the value is not incremented by a segment offset). When the operand is a label name prefixed with “@”, the value pushed depends on what the label points to. The value pushed on the stack is the absolute address represented by the label (i.e., the local label value added to the segment offset).

[0568] 1.26.7. Examples

; constants
.equ SOMECONST, 7
; what follows goes into the data segment
.data
VAR1:
.byte 8
VAR2:
.string “hello\0”
VAR3:
.long 0xFFFCDA07
VAR4:
.long 0
; what follows goes into the code segment
.code
FOO:
PUSH 1
ADD
RET
BAR:
PUSH 2
PUSH @FOO	; push the address of the label FOO
JSR	; jump to the code at label FOO
PUSH SOMECONST	; push the value 7
PUSH @VAR1	; push the addr of VAR1
PUSH VAR1	; push the offset of VAR1 within the data
	segment
PUSH @VAR3	; push the addr of VAR3
PEEK	; push the value of VAR3
PUSH @VAR4	; push the addr of VAR4
POKE	; store the value on top of the stack into
	VAR4
PUSH @VAR2	; push the addr of the string “hello”

[0569] 1.26.8. Command Line Syntax

[0570] In one embodiment, the assembler is a command-line tool that can be invoked with the following syntax: “PktAssembler [options]<input_file_path><output_file_path>”, where the [options] can be: -cs int, -ds int, -xml id, or -h, where “-cs int” is a Code Segment Address Value (default=8), “-ds int” is a Data Segment Address Value (default=4), “-xml id” is used to output a control object as an XML file with the specified ID, and “-h” is used to display help information.

9. Controls

[0571] This section describes illustrative embodiments of control objects. Control objects can be used to represent rules that govern access to content by granting or denying the use of the ContentKey objects they control. They can also be used to represent constraints on the validity of a link object in which they are embedded. They can also be used as standalone program containers that are run on behalf of another entity, such as in agents or delegates. In one embodiment, controls contain metadata and byte-code programs, which implement a specific interaction protocol. The purpose of a Control Protocol is to specify the interaction between the DRM engine and a control program or between a host application and a control program through the DRM engine. This section also describes illustrative actions the application can perform on the content, which action parameters should be supplied to the control program, and how the control program encodes the return status indicating that the requested action can or cannot be performed, as well as parameters that can further describe the return status.

[0572] In this section, the following abbreviations and acronyms are used:

• • [0573] ESB: Extended Status Block
  • [0574] LSB: Least Significant Bit
  • [0575] Byte: 8-bit value, or octet
  • [0576] Byte Code: stream of bytes that encode executable instructions and their operands

[0577] 1.27. Control Programs

[0578] In one embodiment, a control object contains a control program. The control program includes a code module containing byte-code that is executable by a virtual machine, and a list of named routines (e.g., entries in the export table).

[0579] In one embodiment, the set of routines that represent the rules that govern the performance of a certain operation (such as “play”) on a content item is called an ‘action control’. The set of routines that represent validity constraints on a link object is called a “link constraint”. The set of routines that are intended to be executed on behalf of a remote entity (such as during a protocol session with a DRM engine running on a different host) is called an “agent”. The set of routines that are intended to be executed on behalf of another control (such as when a control program uses the System.Host.CallVm system call) is called a “delegate”.

[0580] 1.27.1. Interface to Control Programs

[0581] In one embodiment, control programs are executed by a virtual machine running in a host environment. The host environment can be implemented in any suitable manner; however, for ease of explanation and for purposes of illustration, it will be assumed in the following discussion that the implementation of the virtual machine's host environment can be logically separated into two parts: a host application, and a DRM engine. It will be appreciated, however, that other embodiments may have a different logical separation of functions, which may be equivalent to the logical structure described above.

[0582] As was shown in FIG. 29, in preferred embodiments, the DRM engine 2908 is the logical interface between the host application 2900 and control programs 2906. The host application 2900 makes logical requests to the engine 2908, such as requesting access to a content key for a certain purpose (e.g., to play or render a content stream). In one embodiment, the engine 2908 ensures that the interaction protocol described below is implemented correctly, such as by ensuring that any guarantees regarding a control program's initialization, call sequence, and other interaction details are met.

[0583] When the host application 2900 requests the use of content keys for a set of content IDs, the DRM engine 2908 determines which Control object to use. The Protector objects allow the engine to resolve which ContentKey objects need to be accessed for the requested content IDs. The engine then finds the Controller object that references those ContentKey objects. In one embodiment, a Controller object can reference more than one ContentKey object. This allows multiple ContentKey objects to be governed by the same Control object. When the host application requests access to a content key by invoking an action, it can request content IDs as a group, to the extent that the ContentKey objects that correspond to them are referenced by the same Controller object. In one embodiment, a request to access a group of content keys referenced by more than one controller object is not allowed.

[0584] In one embodiment, the DRM engine follows a convention for mapping actions to routine names. For example, in one embodiment, for each of the routines described below, the name that appears in the Export Table entry in the code module is the respective string shown below in Sections 9.1.4-9.1.7.

[0585] 1.27.1.1. Control Loading

[0586] In one embodiment, before the engine can make calls to control routines, it needs to load the control's code module into the virtual machine. In one embodiment, only one code module per VM is loaded.

[0587] 1.27.1.2. Atomicity

[0588] In one embodiment, the engine ensures that calls to routines within control programs are atomic with respect to the resources it makes available to the routine, such as the object (or “state”) database. Thus, in such an embodiment, the engine needs to ensure that those resources remain unmodified during the execution of any of the routines it calls. This may be done by effectively locking those resources during a routine call, or by preventing multiple VMs to run concurrently. However, the engine need not guarantee that those resources are unmodified across successive routine invocations.

[0589] 1.27.2. Control Protocol

[0590] In one embodiment, the routine naming, the input/output interface, and the data structures for each routine in a code module, together, constitute a Control Protocol. The protocol implemented by a code module is signaled in the Control object's “protocol” field. The illustrative Control Protocol described below will be called the Standard Control Protocol, and its identifier (the value of the ‘protocol’ field) is “http://www.octopus-drm.com/specs/scp-1_—0”.

[0591] In one embodiment, before the DRM engine loads a code module and calls routines in the control program, it needs to guarantee that the interaction with the control program will be consistent with the specification for the specific protocol id signaled in the protocol field. That includes any guarantee about the features of the virtual machine that need to be implemented, guarantees about the size of the address space available to the control program, and the like.

[0592] It is possible for control protocols, such as the Standard Control Protocol, to evolve over time without having to create a new protocol specification. As long as the changes made to the protocol are consistent with previous revisions of the specification, and as long as existing implementations of the DRM engine, as well as existing control programs that comply with that protocol, continue to perform according to the specification, then the changes are deemed compatible. Such changes may include, for instance, new action types.

[0593] 1.27.3. Byte Code Type

[0594] In the illustrative embodiment described above involving the Standard Control Protocol, the type of the byte-code module is “Plankton byte-code module version 1.0”. In this example embodiment, the value for the “type” field of the Control object is “http://www.octopus-drm.com/specs/pkcm-1_—0”.

[0595] 1.27.4. General Control Routines

[0596] General routines are routines that are applicable to the control as a whole, and are not specific to a given action or link constraint. The following general control routines are used in one illustrative embodiment:

[0597] 1.27.4.1. Control.Init

[0598] This routine is optional (i.e., it is not required in all controls). If this routine is used, the engine calls it once before any other control routine is called. The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success, or a negative error code on failure. In one embodiment, if the ResultCode is not 0, the engine aborts the current control operation and does not make any further calls to routines for this control.

[0599] 1.27.4.2. Control.Describe

[0600] This routine is optional. The routine is called when the application requests a description of the meaning of the rules represented by the control program in general (i.e. not for a specific action). The routine has no inputs, and returns a ResultCode and a StatusBlockPointer to the top of the stack as outputs, where the ResultCode is an integer value (0 if the routine completed successfully, or a negative error code otherwise), and where the StatusBlockPointer is the address of a standard ExtendedStatusBlock. The ExtendedStatusBlock contains information that an application can interpret and use to provide information to the user regarding the meaning of the rules represented by the control program.

[0601] 1.27.4.3. Control.Release

[0602] This routine is optional. If this routine exists, the DRM engine calls it once after it no longer needs to call any other routine for the control. No other routine will be called for the control unless a new use of the control is initiated (in which case, the Control.Init routine will be called again). The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success, or a negative error code on failure.

[0603] 1.27.5. Action Routines

[0604] Each possible action has a name (e.g., play, transfer, export, etc.). In one illustrative embodiment, for a given action <Action>, the following routine names are defined (where “<Action>” denotes the actual name of the action (e.g., “play”, “transfer”, “export”, etc.)):

[0605] 1.27.5.1. Control.Actions.<Action>.Init

[0606] This routine is optional. If it exists, the engine calls it once before any other routine is called for this action. The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success, or a negative error code on failure. In one embodiment, if ResultCode is not 0, the engine aborts the current action and does not make any further calls to routines for this action in this control.

[0607] 1.27.5.2. Control.Actions.<Action>.Check

[0608] In the illustrate embodiment being discussed, this routine is required, and is called to check, without actually performing a given action, what the return status would be if the Perform routine were to be called for that action. It is important for this routine not to have any side effects. Note that if the Perform routine also has no side effects, the Check and Perform entries in the control's Entries Table can point to the same routine. This routine has the same inputs and outputs as the Perform routine described below.

[0609] 1.27.5.3. Control.Actions.<Action>.Perform

[0610] In one embodiment, this routine is required, and is called when the application is about to perform the action. The routine has no inputs, and returns a ResultCode and a StatusBlockPointer to the top of the stack as outputs, where the ResultCode is an integer value (0 if the routine completed successfully, or a negative error code otherwise), and where the StatusBlockPointer is the address of a standard ExtendedStatusBlock. Note that in one embodiment a success ResultCode (i.e., 0) does not mean that the request was granted. It only means that the routine was able to run without error. It is the ExtendedStatusBlock that indicates whether the request was granted or denied. However, if the ResultCode indicates a failure, the host application proceeds as if the request was denied. For example, in one embodiment the StatusBlock's category should be ACTION_DENIED, or the returned ExtendedStatusBlock is rejected, and the host application aborts the action.

[0611] When an action is performed, only the Perform routine needs to be called. The engine does not need to call the Check routine beforehand. An implementation of the Perform routine can call the Check routine internally if it chooses to do so, but should not assume that the system will have called the Check routine beforehand.

[0612] 1.27.5.4. Control.Actions.<Action>.Describe

[0613] This routine is optional, and is called when an application requests a description of the meaning of the rules and conditions represented by the control program for the given action. The routine has no inputs, and returns a ResultCode and a StatusBlockPointer to the top of the stack as outputs, where the ResultCode is an integer value (0 if the routine completed successfully, or a negative error code otherwise), and where the StatusBlockPointer is the address of a standard ExtendedStatusBlock.

[0614] 1.27.5.5. Control.Actions.<Action>.Release

[0615] This routine is optional. If it exists, it is called once after the DRM engine no longer needs to call any other routines for the given action. No other routine are called for the given action unless a new use of the action is initiated (in which case, the Init routine will be called again). The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success and a negative error code on failure. If the ResultCode is not 0, the engine does not make any further calls to routines for the given action

[0616] 1.27.6. Link Constraint Routines

[0617] In one embodiment, when a link object has an embedded control, the DRM engine calls the link constraint routines in that control to verify the validity of the link object. The following link constraint routines are used in one illustrative embodiment:

[0618] 1.27.6.1. Control.Link.Constraint.Init

[0619] This routine is optional, and, if it exists, is called exactly once before any other routine is called for the given link constraint. The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success and a negative error code on failure. If the ResultCode is not 0, the engine deems the validity constraint for the link object to be unsatisfied, and avoids making further calls to routines for the link control.

[0620] 1.27.6.2. Control.Link.Constraint.Check

[0621] In the illustrative embodiment being discussed, this routine is required, and is called to check if the validity constraint for a given link is satisfied. The routine has no inputs, and returns a ResultCode and a StatusBlockPointer to the top of the stack as outputs, where the ResultCode is an integer value (0 if the routine completed successfully, or a negative error code otherwise), and where the StatusBlockPointer is the address of a standard ExtendedStatusBlock. If the ResultCode is not 0, the engine deems the validity constraint for the link object to be unsatisfied, and avoids making further calls to routines for the link control. Even if the ResultCode is 0 (success), this does not mean that the constraint has been satisfied; it only means that the routine was able to run without error. It is the StatusBlock that indicates whether the constraint is satisfied or not.

[0622] 1.27.6.3. Control.Link.Constraint.Describe

[0623] This routine is optional, and is called when the application requests a description of the meaning of the constraint represented by the control program for a given link. The routine has no inputs, and returns a ResultCode and a StatusBlockPointer to the top of the stack as outputs, where the ResultCode is an integer value (0 if the routine completed successfully, or a negative error code otherwise), and where the StatusBlockPointer is the address of a standard ExtendedStatusBlock.

[0624] 1.27.6.4. Control.Link.Constraint.Release

[0625] This routine is optional, and, if it exists, is called by the engine once after the engine no longer needs to call any other routine for the given constraint. The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success and a negative error code on failure. In the embodiment being discussed, after calling this routine, no other routine can be called for the given constraint unless a new cycle is initiated (in which case, the Init routine is called again). Similarly, if the ResultCode is not 0, the engine does not make further calls to routines for the given link constraint.

[0626] 1.27.7. Agent Routines

[0627] In one embodiment, an agent is a control object that is designed to run on behalf of an entity. Agents are typically used in the context of a service interaction between two endpoints, where one endpoint needs to execute some virtual machine code within the context of the second endpoint, and possibly obtain the result of that execution. In one embodiment, a control can contain multiple agents, and each agent can contain any number of routines that can be executed; however, in practice, agents typically have a single routine.

[0628] In one illustrative embodiment, the following entry points are defined for agents, where <Agent> is a name string that refers to the actual name of an agent.

[0629] 1.27.7.1. Control.Agents.<Agent>.Init

[0630] This routine is optional, and, if it exists, the engine calls it once before any other routine is called for the given agent. The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success and a negative error code on failure.

[0631] 1.27.7.2. Control.Agents.<Agent>.Run

[0632] In the illustrative embodiment under discussion, this routine is required, and is the main routine of the agent. The routine has no inputs, and returns a ResultCode, a ReturnBlockAddress, and a ReturnBlockSize to the top of the stack as outputs. The ResultCode is an integer value (0 if the routine completed successfully, or a negative error code otherwise), the ReturnBlockAddress is the address of a block of memory that contains data that the agent code is expected to return to the caller (if the routine does not need to return anything, the address is 0), and the ReturnBlockSize is the size in bytes of the block of memory at the ReturnBlockAddress. In one embodiment, if ReturnBlockAddress is 0, the value of ReturnBlockSize is also 0.

[0633] 1.27.7.3. Control.Agents.<Agent>.Describe

[0634] This routine is optional, and is called when an application request a description of a given agent. The routine has no inputs, and returns a ResultCode and a StatusBlockPointer to the top of the stack as outputs, where the ResultCode is an integer value (0 if the routine completed successfully, or a negative error code otherwise), and where the StatusBlockPointer is the address of a standard ExtendedStatusBlock.

[0635] 1.27.7.4. Control.Agents.<Agent>.Release

[0636] This routine is optional, and, if it exists, the engine calls it once after the engine no longer needs to call any other routines for this agent. No other routine will be called for this agent unless a new cycle is initiated (in which case, the Init routine will be called again). The routine has no inputs, and returns a ResultCode to the top of the stack as an output. The ResultCode is 0 on success and a negative error code on failure.

[0637] 1.28. Extended Status Blocks

[0638] The following example definitions are applicable to the ExtendedStatusBlock data structures returned by illustrative embodiments of several of the routines described above. Examples of ExtendedStatusBlock data structures are described in connection with the description of the virtual machine.

[0639] In one embodiment, there are no global ExtendedStatusBlock flags. In this embodiment, control programs set the GlobalFlag field of the ExtendedStatuBlock to 0.

[0640] 1.28.1. Categories

[0641] The following paragraphs define values for the Category field of ExtendedStatusBlocks in accordance with one embodiment. In one embodiment, none of these categories have sub-categories, and thus the value of the SubCategory field of the ExtendedStatusBlocks is set to 0.

[0642] In one embodiment, the following category codes are defined:

[0643] 1.28.1.1. Actions Check and Perform Routines

Value	Name	Description
0	ACTION_GRANTED	The application is authorized to use the content
		keys controlled by the control program for the purpose of
		the requested action.
		The parameter list of the returned
		ExtendedStatusBlock should not contain any of
		the constraint parameters, but may contain
		obligation and/or callback parameters.
1	ACTION_DENIED	The application is not authorized to use the content
		keys controlled by the control program for the
		purpose of the requested action.
		When an action is denied, the control program
		should include in the parameter list of the returned
		ExtendedStatusBlock one or more of the
		constraints that were not met and caused the action
		to be denied (the constraints that were not
		evaluated and the constraints that did not cause the
		action to fail should be omitted).
		In one embodiment, the parameter list of the
		returned ExtendedStatusBlock must not contain
		any obligation or callback parameter.

[0644] In one embodiment, in the context of ExtendedStatusBlock parameters returned by action routines, a constraint means a condition that is required to be true or a criterion that is required to be met in order for the result of the routine to return an ExtendedStatusBlock with the category ACTION_GRANTED.

[0645] In one embodiment, values for the LocalFlags field common to both categories described above include:

Bit
Index (0 is LSB)	Name	Description
0	OBLIGATION_NOTICE	The parameter list contains one or
		more parameters that are related to
		obligations
1	CALLBACK_NOTICE	The parameter list contains one or
		more parameters that are related to
		callbacks
2	GENERIC_CONSTRAINT	The parameter list contains one or
		more parameters that are related to
		generic constraints
3	TEMPORAL_CONSTRAINT	The parameter list contains one or
		more parameters that are related to
		temporal constraints
4	SPATIAL_CONSTRAINT	The parameter list contains one or
		more parameters that are related to
		spatial constraints
5	GROUP_CONSTRAINT	The parameter list contains one or
		more parameters that are related to
		group constraints
6	DEVICE_CONSTRAINT	The parameter list contains one or
		more parameters that are related to
		device constraints
7	COUNTER_CONSTRAINT	The parameter list contains one or
		more parameters that are related to
		counter constraints

[0646] In the table shown above, the parameter list that is referred to is the “Parameters” field of the ExtendedStatusBlock data structure.

[0647] 1.28.1.2. Describe Routine Category Codes

[0648] In one embodiment, no category codes are defined for Describe routines. In one embodiment, the same local flags as the ones defined for Action routines apply to Describe routines, and Describe routines should include in their retuned ExtendedStatusBlock a parameter named ‘Description’ as specified below. In one embodiment, Describe routines do not contain in their retuned ExtendedStatusBlock any obligation or callback parameters; however, Describe routines should include in their returned ExtendedStatusBlock parameters that describe some or all of the constraints that are applicable for the corresponding action or link constraint.

[0649] 1.28.1.3. Link Constraint Routine Category Codes

Value	Name	Description
0	LINK_VALID	The link constrained by this control program is
		valid.
		The parameter list of the returned ESB should
		not contain any of the constraint parameters,
		and, in one embodiment, must not contain
		obligation or callback parameters
1	LINK_INVALID	The link constrained by this control program is
		invalid.
		When a link is invalid, the control program
		should include in the parameter list of the
		returned ESB one or more of the
		constraints that were not met and caused
		the link to be invalid (the constraints
		that were not evaluated and the
		constraints that did not cause the action
		to fail should be omitted). In one
		embodiment, the parameter list of the
		returned ESB must not contain any
		obligation or callback parameter.

[0650] In one embodiment, the same local flags as the ones defined for Action routines apply for each of these categories.

[0651] In one embodiment, in the context of ExtendedStatusBlock parameters returned by link constraint routines, a constraint means a condition that is required to be true or a criterion that is required to be met in order for the result of the routine to return an ExtendedStatusBlock with the category LINK_VALID.

[0652] 1.28.2. Cache Durations

[0653] The CacheDuration field of an ExtendedStatusBlock is an indication of the validity period of the information encoded in the ExtendedStatusBlock. When an ExtendedStatusBlock has a non-zero validity period, it means that the ExtendedStatusBlock can be stored in a cache, and that during that period of time a call to the exact same routine call with the same parameters would return the same ExtendedStatusBlock, so the cached value may be returned to the host application instead of calling the routine.

[0654] 1.28.3. Parameters

[0655] Some parameters are used to convey detailed information about the return status, as well as variable bindings for template processing (see Section 9.4).

[0656] In one embodiment, except for obligations and callbacks, all the constraints described here are strictly for the purpose of helping the host application classify and display, not for enforcement of the usage rules. The enforcement of the rules is the responsibility of the control program.

[0657] In one embodiment, the parameters defined in the following section are encoded either as a ParameterBlock, if no parameter flags are applicable, or as an ExtendedParameterBlock, of one or more flags are applicable. Representative flags are described below:

[0658] 1.28.3.1. Description

[0659] Parameter Name: Description

[0660] Parameter Type: ValueList

[0661] Description: List of description parameters. Each value in the list is of type Parameter or Extended Parameter. In one embodiment, the following parameters are defined: Default, Short and Long. Each of them, if present has for a value the ID of one of the control's resources. That resource should contain a textual payload, or a template payload. If the resource is a template, it is processed to obtain a textual description of the result (either a description of the entire control program, or of a specific action). The template is processed using as variable bindings the other parameters of the list in which the ‘Description’ parameter appears.

[0662] In one embodiment, the ‘Short’ and ‘Long’ descriptions can only be included if a ‘Default’ description is also included.

	Name	Type	Description
	Default	Resource	Id of the resource that contains
			the normal description text or template
	Short	Resource	Id of the resource that contains
			the short description text or template
	Long	Resource	Id of the resource that contains
			the long description text or template

[0663] 1.28.3.2. Constraints

[0664] In one embodiment, constraint parameters are grouped in lists that contain constraints of similar types. In one embodiment, standard constraints are defined for some of the types. In one embodiment, controls may return constraint parameters that are not included in the set of standard constraints, provided that the name of the constraint parameters be a URN in a namespace that guarantees the uniqueness of that name. This may include vendor-specific constraints, or constraints defined in other specifications.

[0665] 1.28.3.2.1. Generic Constraints

[0666] Parameter Name: GenericConstraints

[0667] Parameter Type: ValueList

[0668] Description: List of generic constraints that may be applicable. Each value in the list is of type Parameter or ExtendedParameter.

[0669] In one embodiment, generic constraints are constraints that do not belong to any of the other constraint types defined in this section. In one embodiment, no generic constraint parameters are defined.

[0670] 1.28.3.2.2. Temporal Constraints

[0671] Parameter Name: TemporalConstraints

[0672] Parameter Type: ValueList

[0673] Description: List of temporal constraints that may be applicable. Each value in the list is of type Parameter or Extended Parameter. Temporal constraints are constraints that are related to time, date, duration, and/or the like. In one embodiment, the following temporal constraint parameters are defined:

Name	Type	Description
NotBefore	Date	Date before which the action is denied
NotAfter	Date	Date after which the action is denied
NotDuring	ValueList	List of 2 values of type Date. The first
		value is the start of the period, and the
		second is the end of the period that
		is excluded.
NotLongerThan	Integer	Max number of seconds after first use.
		In one embodiment, this value is unsigned.
NotMoreThan	Integer	Max number of seconds of accumulated
		use time. In one embodiment, this value
		is unsigned.

[0674] 1.28.3.2.3. Spatial Constraints

[0675] Parameter Name: SpatialConstraints

[0676] Parameter Type: ValueList

[0677] Description: List of spatial constraints that may be applicable. In one embodiment, each value in the list is of type Parameter or ExtendedParameter. Spatial constraints are constraints that are related to physical locations. In one embodiment, no standard spatial constraints are defined.

[0678] 1.28.3.2.4. Group Constraints

[0679] Parameter Name: GroupConstraints

[0680] Parameter Type: ValueList

[0681] Description: List of group constraints that may be applicable. Each value in the list is of type Parameter or Extended Parameter. Group constraints are constraints that are related to groups, group membership, identity groups, and/or the like. In one embodiment, the following parameters are defined:

Name	Type	Description
MembershipRequired	Resource	Id of the resource that contains the
		text or template for the name or
		identifier of a group of which a
		membership is required
IdentityRequired	Resource	Id of the resource that contains the
		text or template for the name or
		identifier of an individual

[0682] 1.28.3.2.5. Device Constraints

[0683] Parameter Name: DeviceConstraints

[0684] Parameter Type: ValueList

[0685] Description: List of device constraints that may be applicable. Each value in the list is of type Parameter or Extended Parameter. Device constraints are constraints that are related to characteristics of a device, such as features, attributes, names, identifiers, and/or the like. In one embodiment, the following parameters are defined:

Name	Type	Description
DeviceTypeRequired	Resource	Id of the resource that contains the
		text or template for the type of host
		device that is required
DeviceFeatureRequired	Resource	Id of the resource that contains
		the text or template for name of
		feature that the host device
		must have
DeviceIdRequired	String	Id that the device is required to
		have. This Id may be any string
		that can be used to identify the
		device (e.g., device name,
		device serial number, a node
		id, and/or the like).

[0686] 1.28.3.2.6. Counter Constraints

[0687] Parameter Name: CounterConstraints

[0688] Parameter Type: ValueList

[0689] Description: List of counter constraints that may be applicable. Each value in the list is of type Parameter or ExtendedParameter. Counter constraints are constraints that are related to counted values, such as play counts, accumulated counts, and/or the like. In one embodiment, no standard counter constraints are defined.

[0690] 1.28.3.3. Parameter Flags

[0691] In one embodiment, the following flags may be used for all the parameters described in Section 9.2.3, when they are encoded as an ExtendedStatusBlock:

Bit
Index
(0 is
LSB)	Name	Description
0	CRITICAL	The semantics associated with this
		parameter need to be understood by the
		host application. If they are not, the
		entire ExtendedStatusBlock should
		be treated as not understood
		and rejected.
		In one embodiment, this flag should not
		be used for parameters that are
		descriptive in nature.
1	HUMAN_READABLE	This parameter represents a value
		whose name and value are suitable to
		display in a textual or graphical user
		interface. Any parameter that does
		not have this flag set should be reserved
		for the host application, and not be
		shown to a user. For parameter
		values of type Resource, it
		is not the resource ID, but the resource
		data payload referenced by the ID,
		that is human-readable.

[0692] 1.29. Obligations and Callbacks

[0693] In one embodiment, certain actions, when granted, require further participation from the host application. Obligations represent operations that need to be performed by the host application upon or after the use of the content key it is requesting. Callbacks represent calls to one or more of the control program routines that need to be performed by the host application upon or after the use of the content key they are requesting.

[0694] In one embodiment, if an application encounters any critical obligation or callback that it does not support, or does not understand (for example because the obligation type may have been defined after the application was implemented), it must refuse to continue the action for which this obligation or callback parameter was returned. In one embodiment, a critical obligation or callback is indicated by setting the CRITICAL parameter flag for the parameter that describes it.

[0695] If a control has side effects (such as decrementing a play count, for example), it should use the OnAccept callback to require the host application to call a certain routine if it is able to understand and comply with all critical obligations and callbacks. The side effect should happen in the callback routine. In one example embodiment, implementations are required to understand and implement the OnAccept callback, since it can be useful in preventing side effects (e.g., updates to the state database) from occuring prematurely (e.g., before the host application determines that it is unable to comply with a given critical obligation or callback and needs to terminate performance of an action), thus providing a measure of transactional atomicity.

[0696] 1.29.1. Parameters

[0697] The following parameters define several types of obligations and callbacks that can be returned in ExtendedStatusBlock data structures.

[0698] 1.29.1.1. Obligations

[0699] Parameter Name: Obligations

[0700] Parameter Type: ValueList

[0701] Description: List of obligation parameters. Each value in the list is of type Parameter or Extended Parameter. In one embodiment, the following obligation parameters are defined:

Name	Type	Description
RunAgentOnPeer	ValueList	The host application needs to send an
		agent control to run on a peer of the
		currently running protocol session
	Type	Description
	String	Id of the Control that contains
		the agent to run.
	String	Name of the agent to run.
	Integer	Instance Id. This value is used
		to uniquely identify this agent
		obligation instance. This id
		will also allow the system to
		correlate this agent obligation
		with an OnAgentCompletion
		callback parameter.
	String	Context Id. This Id will be
		visible to the running agent on
		the peer under the agent's
		session context Host Object
		path:
		Octopus/Agent/Parameters/
		Session/ContextId.
	ValueList	List of values of type
		Parameter. All those
		parameters will be
		visible to the agent as input
		parameters.

[0702] 1.29.1.2. Callbacks

[0703] Parameter Name: Callbacks

[0704] Parameter Type: ValueList

[0705] Description: List of callback parameters. Each value in the list is of type Parameter or Extended Parameter. In one embodiment, the following callbacks parameters are defined:

Name	Type	Description
OnAccept	Callback	The host application must call back if it is able to
		understand all the critical obligations and callback
		parameters contained in this ESB.
		In one embodiment, there can be at most one
		OnAccept callback parameter in a list of callback
		parameters. If other callback parameters are
		specified in the list, the OnAccept is executed first.
OnTime	ValueList	The host application must call back after the
		specified date/time.
	Type	Description
	Date	The date after which the host
		application needs to perform the
		callback.
	Callback	Routine to call back, and associated
		cookie.
OnTimeElapsed	ValueList	The host application must call back after the
		specified duration has elapsed (the counting starts
		when the host application actually performs the
		action for which the permission that was granted).
	Type	Description
	Integer	Number of seconds. The value is
		unsigned.
	Callback	Routine to call back, and associated
		cookie.
OnEvent	ValueList	The host application must call back when a certain
		event occurs.
	Type	Description
	String	Event Name
	Integer	Event Flags (the integer valus is
		interpreted as a bit-field)
	Integer	Event Parameter
	Callback	Routine to call back, and associated
		cookie.
		See the paragraph about events for more details
		about the events.
OnAgentCompletion	ValueList	The host application must call back when an agent
		specified in one of the obligation parameters has
		completed, or failed to run.
	Type	Description
	Integer	Agent instance id.
		The instance id specified in an agent
		obligation.
	Callback	Routine to call back, and associated
		cookie.
	When calling back, the host application must
	provide the following ArgumentsBlock:
	Type	Encoding	Description
	32-bit	4 bytes in	Completion status
	integer	big-endian	code.
		order
	32-bit	4 bytes in	Agent result code
	integer	big-endian
		order
	8-bit byte	Byte	Agent ReturnBlock
	array	sequence
	The completion status code value is 0 if the agent
	was able to run or a negative error code if it was
	not.
	The agent ReturnBlock is the data returned by the
	agent. This is omitted if the agent was unable to run
	(the Completion status code is not 0).

[0706] In one embodiment, the ‘Callback’ type mentioned in the table above is a ValueListBlock with three ValueBlock elements:

Value
Type	Description
Integer	ID of the callback type. In one embodiment, two types of
	callbacks are defined:
	ID	Description
	RESET = 0	All pending callbacks requests and active
		obligations are cancelled upon calling the
		callback routine. The callback routine
		returns an ESB that indicates if and how the
		application can continue with the current
		operation.
	CONTINUE = 1	The callback routine is called while all
		other pending callback requests and active
		obligations remain in effect. The callback
		routine returns a simple result code. The
		application can continue with the current
		operation unless that result code indicates a
		failure.
String	Entry point to call in the code module. In one embodiment,
	this must be one of the entries in the Export Table of the code
	module for the same control as the one containing the routine
	that returned the ESB with this parameter.
Integer	Cookie. This value will be passed on the stack to the
	routine that is called.

[0707] 1.29.1.3. Parameter Flags

[0708] In one embodiment, the same parameter flags as defined in the previous section are used. In one embodiment, callbacks and obligations that a caller is required to implement are marked as CRITICAL, so as to avoid giving a host application the choice to ignore these parameters.

[0709] 1.29.2. Events

[0710] In one embodiment, events are specified by name. Depending on the type of event, there may be a set of flags defined that further specify the event. In one embodiment, if no flags are defined for a specific event, the value of the flag field is set to 0. Also, some events may specify that some information be supplied to the callback routine when the event occurs. In one embodiment, if no special information is required from the host application, the host application must call with an empty ArgumentsBlock (see the description of the callback routine interface in section 3.3, below).

[0711] In one embodiment, if the name of an event in a callback parameter marked CRITICAL is not understood or not supported by the host application, the host application must consider this parameter as a not-understood CRITICAL parameter (and the action for which permission was requested must not be performed).

[0712] In one embodiment, the following event names are defined:

		Event
Event Name	Event Flags	Parameter	Description
OnPlay	None	None	The host application must call back
			when the multimedia object starts
			playing.
OnStop	None	None	The host application must call back
			when the multimedia stops playing (or is
			paused)
OnTimecode	None	Presentation	The host application must call back
		time	when the specified presentation time has
		expressed in	been reached or exceeded (during
		number of	normal real-time playback or after a
		seconds since	seek). The origin of the presentation
		the start of	time is when the rendering begins. The
		the	presentation time relates to the source
		presentation	media time, not the wall-clock time
			(e.g., when a presentation is paused, the
			presentation time does not change).
OnSeek	None	None	The host application must call back
			when a direct access to an arbitrary point
			in a multimedia presentation occurs.
			In one embodiment, when calling back,
			the host application must provide the
			following data in a ArgumentsBlock:
	Type	Encoding	Description
	32-bit	4 bytes in	Seek
	unsigned	big-endian	position
	integer	order	offset
	32-bit	4 bytes in	Seek
	unsigned	big-endian	position
	integer	order	range
	The position within the multimedia
	presentation is offset “marks” out of
	range total “marks” in the presentation.
	For instance, for a presentation that is
	327 seconds long, seeking to the 60th
	second can be represented with
	offset = 60, range = 327. It is up to the
	caller to choose the unit that corresponds
	to the measurement of the offset and
	range (it could be a time unit, a byte-size
	unit, or any other unit), provided that the
	“marks” are homogeneously distributed
	over the entire presentation. The value
	of offset must be less than or equal to the
	value of range.

[0713] 1.29.3. Callback Routines

[0714] In one embodiment, callback routines take the same input:

[0715] Input: Top of stack:

	Cookie
	ArgumentsBlockSize
	. . . data . . .

[0716] Cookie: the value of the Cookie field that was specified in the callback parameter.

[0717] ArgumentsBlockSize: number of bytes of data passed on the stack below this parameter. When the routine is called, the stack contains the value ArgumentsBlockSize supplied by the caller, indicating the size of the arguments block at the top, followed by ArgumentsBlockSize bytes of data. In one embodiment, if the size is not a multiple of 4, the data on the stack will be padded with 0-value bytes to ensure that the stack pointer remains a multiple of 4.

[0718] 1.29.3.1. CONTINUE Callbacks

[0719] In one embodiment, callbacks with the type CONTINUE (type ID=0) have the following output:

[0720] Output: Top of stack:

	ResultCode
	. . .

[0721] ResultCode: an integer value. The result value is 0 if the routine was able to execute or a negative error code if an error occurred.

[0722] Description: if the ResultCode indicates that the callback routine was able to run (i.e., the value is 0), the host application can continue the current operation. If the ResultCode indicates that an error occurred, the host application aborts the current operation and cancels all pending callbacks and obligations.

[0723] 1.29.3.2. RESET Callbacks

[0724] When a control routine has specified one or more callbacks of type RESET in the ESB returned from a routine, the host application will call any specified callback routine when the condition for that callback is met. In one embodiment, as soon as the conditions of any of the callbacks are met, the host application needs to:

• • [0725] Cancel all other pending callbacks
  • [0726] Cancel all current obligations
  • [0727] Provide the required parameters (if any) for that callback
  • [0728] Call the specified callback routine.

[0729] The return status from the routine indicates to the host application if it can continue performing the current operation. In one embodiment, if the permission is denied or the routine fails to execute successfully, the host application must abort the performance of the current operation. Similarly, if the permission is granted, the host application must comply with any obligation or callback that may be returned in an ESB, just as if it had called the original Control.Actions.<Action>.Perform routine. Previous obligations or callback specifications are no longer valid.

[0730] In one embodiment, all routines specified as callback entry points for this type of callback have the following output:

[0731] Output: Top of stack:

	ResultCode
	StatusBlockPointer
	. . .

[0732] ResultCode: an integer value. The result value is 0 if the routine was able to execute, or a negative error code if an error occurred.

[0733] StatusBlockPointer: address of a standard ExtendedStatusBlock.

[0734] Description: the return semantics of this routine are equivalent to what is described for the Control.Actions.<Action>.Perform routine.

[0735] 1.30. Metadata Resources

[0736] In one embodiment, control objects can contain metadata resources, which can be referenced from the parameters returned in ExtendedStatusBlock data structures. Resources can be simple text, text templates, or other data types. Each resource is identified by a resource ID, and can contain one or more text strings or encoded data, one for each version in a different language. It is not required that resources be provided for all languages. It is up to the host application to choose which language version is most appropriate for its needs.

Field	Type	Description
Resource
Id	ASCII String	URI (typically a URN referring to
		the Id of an Extension of the
		Control object that contains the
		code module with the routine that is
		currently running)
Type	ASCII String	MIME-type of the resource data as
		described in IETF RFC 2046
Data	List of	List of all the different versions of
	LocalizedData	the resource, for different locales
LocalizedData
Language	ASCII String	Language code as specified in IETF
		RFC 3066
Data	Type depends	The actual data for the resource
	on the specified	(text, etc . . . )
	mime type

[0737] Resources accompany control programs by being included as Extensions in a Control object. The resource Id maps to the Id of an internal extension of the Control object that contains the code module with the routine that is currently running.

[0738] For the purpose of computing the canonical byte sequence for Resource objects, in one embodiment the data structure description is the following:

	class LocalizedData {
	string language
	byte[ ] data;
	}
	class Resource {
	string id
	string type;
	LocalizedData data;
	}

[0739] 1.30.1. Simple Text

[0740] Simple text is specified as MIME-type ‘text’

[0741] 1.30.2. Text Templates

[0742] In addition to the standard text resources, in one embodiment, a text template type is defined. The MIME-type for this is ‘text/vnd.intertrust.octopus-text-template’.

[0743] In one embodiment, a text template contains text characters encoded in UTF-8, as well as named placeholders that are to be replaced by text values obtained from parameters returned in the parameters list, such as that of an ExtendedStatusBlock. The syntax for a placeholder is ‘\PLACEHOLDER\’, where PLACEHOLDER specifies the name of a Parameter Block and an optional formatting hint. In one embodiment, the template processor must replace the entire token ‘\PLACEHOLDER\’ with the formatted representation of the Value field of that Parameter Block, and the formatting of the Value data is specified below in Section 4.2.1.

[0744] In one embodiment, if the character ‘\’ appears in the text outside of a placeholder, it must be encoded as ‘\\’, and all occurrences of ‘\\’ in the text will be reverted to ‘\’ by the template processor.

[0745] The syntax for the placeholder is: FORMAT|NAME, where NAME is the name of a Parameter Block, and FORMAT is the formatting hint to convert the parameter's data into text. If the default formatting rules for the parameter value's data type are sufficient, then the formatting hint can be omitted, and the placeholder is simply NAME.

[0746] 1.30.2.1. Formatting

[0747] 1.30.2.1.1. Default Formatting

[0748] In one embodiment, the default formatting rules for the different value types are:

Type	Formatting
Integer	Text representation of the integer value as a signed
	decimal. The text is composed only of the characters for
	the digits “0” to “9” and the character “—”. If the value is 0,
	the text is the string “0”. If the value is not 0, the text does
	not start with the character “0”. If the value is negative, the
	text starts with the character “—”. If the value is positive, the
	text starts with a non-zero digit character.
Real	Text representation of the floating point value in decimal.
	The integral part of the value is represented using the same rules
	as for Integer values. The decimal separator is represented
	with the host application's preferred decimal separator. The
	factional part of the value consists of up to 6 “0” characters
	followed by up to 3 non-zero digit characters.
String	The string value itself
Date	A human readable representation of the date, according to
	the host's preferred text representation of dates
Parameter	The text “<name>=<value>”, where <name> is the
	parameter name, and <value> is the parameter value
	formatted according to the default formatting rules for its type.
ExtendedParameter	Same as for Parameter
Resource	Text string of the resource's data. In one embodiment, the
	resource referenced by the placeholder must have a MIMI-
	type that is text-based (e.g., text or
	text/vnd.intertrust.octopus-text-template).
ValueList	The text “<value>, <value>, . . . ” with all the values in the
	list formatted according to the default formatting rules for
	their type.

[0749] 1.30.2.1.2. Explicit Formatting

[0750] Explicit format names can be used as the FORMAT part of a placeholder tag. If an unknown FORMAT name is encountered, the template processing engine will use the default formatting rules.

Name	Formatting
Hex	Hexadecimal representation of an integer value interpreted as
	unsigned. In one embodiment, this formatting hint should be
	ignored for data types that are not integers.

[0751] 1.31. Context Objects

[0752] In one embodiment, when a control routine is executing, it has access to a number of context objects through the use of the System.Host.GetObject system call.

[0753] 1.31.1. General Context

[0754] In one embodiment, the following context is present for running controls.

Name	Type	Description
Octopus/Personality/Id	String	ID of the current
		personality Node
Octopus/Personality/Attributes	Container of	Attributes of the current
	Attributes	personality Node

[0755] 1.31.2. Runtime Context

[0756] In one embodiment, the following context is present for all controls that are running in a VM that has been created using the System.Host.SpawnVm system call. In one embodiment, this context must be non-existent or an empty container for controls that are running in a VM that was not created using System.Host.SpawnVm.

Name	Type	Description
Octopus/Runtime/Parent/Id	Container	The identity under which the
	of unnamed	caller of the system call is
	String	running.
	objects

[0757] 1.31.3. Control Context

[0758] In one embodiment, the following context is present whenever a routine of a control is running:

Name	Type	Description
Octopus/Control/Id	String	Id of the running control
Octopus/Control/	Container	Attributes of the running control.
Attributes		This object may be omitted if the
		control has no attributes.

[0759] 1.31.4. Controller Context

[0760] In one embodiment, the following context is present whenever a routine of a control is running and the control was pointed to by a controller object (e.g., when accessing a ContentKey object in order to consume protected content).

Name	Type	Description
Octopus/Controller/Id	String	Id of the Controller that points to the
		currently running control
Octopus/Controller/	Container	Attributes of the Controller pointing
Attributes		to the currently running control. This
		object may be omitted if the
		controller has no attributes.

[0761] In embodiments where a host application is allowed to only group content keys that are controlled by a single controller object, for a given action, there will be only one applicable controller object.

[0762] 1.31.5. Action Context

[0763] In one embodiment, the following context is present whenever a control is called for the purpose of controlling an Action.

Name	Type	Description
Octopus/Action/	Container	Array of Name/Value pairs
Parameters		representing the parameters that are
		relevant for the current action, if any.
		In one embodiment, each action type
		defines a list of optional and required
		parameters. This container may be
		omitted if the action has no parameters.

[0764] 1.31.6. Link Context

[0765] In one embodiment, the following context is present whenever a control is called for the purpose of limiting the validity of a link object (e.g., a control object embedded in a link object):

Name	Type	Description
Octopus/Link/Id	String	Id of the Link object
Octopus/Link/Attributes	Container	Attributes of the Link object that
		contains the running control. This
		object may be omitted if the link
		has no attributes.

[0766]

Name	Type	Description
Octopus/Agent/Parameters	Container	Array of Name/Value parameter
		pairs representing the input
		parameters for the agent.
Octopus/Agent/Session/	String	Identifier for the session context
ContextId		in which the agent is running.

[0767] The Parameter and Session containers are normally used to allow the protocols that require one entity to send and run an agent on another entity to specify which input parameters to pass to the agent, and which session context objects the host needs to set under certain conditions. The presence or absence of certain session context objects may allow the agent code to decide whether it is running as part of the protocol it was designed to support, or if it is running out of context, in which case it may refuse to run. For example, an agent whose purpose is to create a state object on the host on which it runs may refuse to run unless it is being executed during a specific protocol interaction.

[0768] 1.32. Actions

[0769] In one embodiment, each action has a name and a list of parameters. In one embodiment, some parameters are required—the application must provide them when performing this action—and some are optional—the application may provide them or may omit them.

[0770] In one embodiment, the following standard actions are defined: