System and method for efficiently managing data transports

[0001] 1. Field of the Invention

[0002] This invention relates generally to the field of data processing systems. More particularly, the invention relates to a system and method for prioritizing data transports on a data processing device.

[0003] 2. Description of the Related Art

[0004] A typical layered network architecture is illustrated in FIG. 1. Each layer within the architecture performs a specific function to reliably transmit data from a source node 195 (e.g., a client computer) to a destination node 196 (e.g., a network server). For example, when an application 190 has data to transmit to another application 191, the data is processed, in succession by a transport layer 180, a network layer 170 and a data-link layer 160 before being transmitted over the actual physical connection 165 between the two nodes. At the receiving node 196, the data is then processed in reverse order, by the data-link layer 161, the network layer 171, and the transport layer 185 before being handed off to the receiving application 191.

[0005] The descriptions below assume that the reader has at least a basic understanding of the functions of each of the network layers. For those interested, a detailed description of the network layers defined by the ISO Open Systems Interconnection model can be found in DILIP C. NAIK, INTERNET STANDARDS AND PROTOCOLS (1998) (see, e.g., Chapter 1, pages 3–11).

[0006] The well known TCP/IP protocol (“Transmission Control Protocol/Internet Protocol”) operates at the transport and network layers, respectively. The TCP transport layer is responsible for ensuring that all data associated with a particular data transmission arrives reliably and in the correct order at its destination. Specifically, in order to ensure reliable data transmission, a virtual connection 187 (also referred to herein as a “socket connection”) is established between a TCP socket 182 opened at the destination transport layer 185 and a TCP socket 181 opened at the source transport layer 180.

[0007] The TCP sockets 181, 182 perform flow control to ensure that the data transmitted form the source node is provided to the receiving node at an acceptable data rate. Specifically, a “window” is established defining the amount of outstanding data a source node 195 can send before it receives an acknowledgment back from the receiving node 196 (i.e., indicating that it has successfully received all or a portion of the data).

[0008] For example if a pair of nodes 195, 196 are initially communicating over TCP connection that has a TCP window size of 64 KB (kilobytes), the transmitting socket 181 can only send 64 KB of data and then it must stop and wait for an acknowledgment from the receiving socket 182 that some or all of the data has been received. If the receiving socket 182 acknowledges that all of the data has been received then the transmitting socket 181 is free to transmit another 64 KB. If, however, the transmitting socket 181 receives an acknowledgment from the receiver that it only received the first 32 KB (which could happen, for example, if the second 32 KB was still in transit or was lost), then the transmitting socket 181 will only send another 32 KB, since it cannot have more than 64 KB of unacknowledged data outstanding (i.e., the second 32 KB of data plus the third).

[0009] Thus, the TCP window throttles the transmission speed based on how quickly the receiving application can process it. The TCP window is typically defined by a 16-bit TCP header field. As such, the largest window that can be used for a standard TCP connection is 2¹⁶ KB (64 KB).

[0010] A client may concurrently have several different socket connections open with a server or client, or with several different servers/clients. Each socket connection may not be utilized in the same manner, however. For example, the user may be interactively browsing web pages via one socket connection while receiving an automated software upgrade or e-mail message over another socket connection. This may result in a degradation of the interactive user experience, particularly on networks which allocate a relatively small amount of bandwidth per device (e.g., wireless networks such as Cellular Digital Packet Data and ARDIS networks). Under these circumstances, it would be useful to have the ability to prioritize the socket connections such that the interactive connections are provided with a relatively larger amount of bandwidth than the non-interactive connections.

[0011] A system is described for allocating bandwidth comprising: a transport layer to support a first socket connection for a first application and a second socket connection for a second application; and a socket prioritization module to allocate relatively more bandwidth to the first socket connection relative to the second socket connection, wherein allocation of bandwidth is based on one or more characteristics of the first application and/or the second application.

[0020] Described below is a system and method for processing code words. Throughout the description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced without some of these specific details. In other instances, well-known structures and devices are shown in block diagram form to avoid obscuring the underlying principles of the present invention.

[0021] One embodiment of the invention is employed on a wireless data processing device which operates in conjunction with a data processing service such as that described in co-pending application entitled NETWORK PORTAL SYSTEM, APPARATUS AND METHOD, Ser. No. 09/714,897, Filed Nov. 15, 2000 (hereinafter “Network Portal Application”), which is assigned to the assignee of the present application and which is incorporated herein by reference. Certain aspects of this data processing device and data processing service will now be described, followed by a detailed description of an advanced system and method for processing data object identification codes. As an initial matter, however, it should be noted that the underlying principles of the invention may be implemented on virtually any type of network and between virtually any types of network nodes (e.g., networked desktop computers, networked personal digital assistants, . . . etc).

[0022] FIG. 2 illustrates one embodiment of a data processing service 100 comprised of one or more servers. The data processing service provides a portal through which data processing devices 110 may access content (e.g., Web pages, multimedia content, e-mail, instant messages . . . etc) from external Internet sites 130. In one embodiment, the service 100 converts standard applications and data into a format which each wireless data processing device 110 can properly interpret. Thus, as illustrated in FIG. 2, one embodiment of the service 100 includes a content conversion module 120 for processing requests for Internet content 140. More particularly, the service 100 acts as a proxy for the data processing device 110, forwarding Internet requests 140, 141 to the appropriate Internet site 130 on behalf of the data processing device 110, receiving responses from the Internet site 130 in a standard Internet format (e.g., Web pages with embedded audio/video and graphical content, e-mail messages with attachments, . . . etc), and converting the standard Internet responses 142 into a format 143 which the data processing device 110 can process (e.g., bytecodes as described in the Network Portal Application).

[0023] For example, the conversion module 120 may include a hypertext markup language (“HTML”) rendering module (not shown) for interpreting HTML code and downloading any embedded content in the HTML code (e.g., graphics, video, sound, . . . . etc) to the service 100. The conversion module 120 may then combine the HTML code and embedded content and generate a set of bytecodes for accurately reproducing the requested content on the data processing device 110. As described above, in one embodiment, the bytecodes may be Java bytecodes/applets. However, the conversion module 120 may generate various other types of interpreted and/or non-interpreted code, depending on the particular type of data processing device 110 being used (e.g., one with an interpreter module or one without).

[0024] Because one embodiment of the service 100 maintains an intimate knowledge of the capabilities/configuration of each data processing device 110 (e.g., screen size, graphics/audio capabilities, available memory, processing power, user preferences, . . . etc) it can reconstruct the requested Internet content accurately, while at the same time minimizing the bandwidth required to transmit the content to the device 110. For example, the conversion module 120 may perform pre-scaling and color depth adjustments to the requested content so that it will be rendered properly within the data processing device's 110's display. In making these calculations, the conversion may factor in the memory and processing power available on the data processing device 110. In addition, the conversion module 120 may compress the requested content using a variety of compression techniques, and thereby preserve network bandwidth.

[0025] In one embodiment, the conversion module 120 will simply discard Internet content which either cannot be reproduced on the data processing device 110, or which the user has indicated that he/she does not want to be reproduced on the wireless device. For example, a user may indicate that he/she does not want sounds to be generated on the data processing device 110 or that he/she does not want advertisements transmitted to the data processing device 110. The conversion module 120 will then remove any sounds or advertisements embedded in the requested Web page (or other requested Internet content). Because HTML rendering and other advanced processing of Internet content/data is offloaded to the service 100 as described above, the data processing device 110 can be manufactured using a relatively low power microprocessor or microcontroller, thereby lowering the cost of manufacture and/or the energy consumed by the device 110.

[0026] In one embodiment, when a particular Web page or other Internet object has been converted into a format suitable for rendering/execution on the data processing device 110 the formatted page/object may be stored locally on a cache 125 maintained at the service 100. The next time the content is requested, the conversion module 120 may simply read the previously-generated code from the local cache 125 (i.e., it will no longer need to retrieve the content from remote locations to reconstruct the code).

[0027] One embodiment of a system for prioritizing data transports (e.g., socket connections) is illustrated generally in FIG. 3. Two applications 390 and 391 executed on the data processing device 110 are shown communicating with two service modules 392 and 393, respectively, executed on the service 110. One socket connection, represented by sockets 310 and 312 in transport layers 305 and 306, respectively, provides for reliable data communication between application 390 and service 393. Similarly, a socket connection supported by sockets 311 and 313 provides for reliable data communication between application 391 and service 392.

[0028] In one embodiment, the transport layers 305 and 306 provide all, or a subset of, the standard TCP transport functions. In addition, the embodiment illustrated in FIG. 3 includes socket prioritization logic 330 and 331 which prioritizes socket connections based on a one or more characteristics of the applications for which they are generated. For example, sockets for interactive applications such as Web browsing and instant messaging may be assigned relatively high priorities whereas sockets for non-interactive applications such as automated software upgrades may be assigned relatively low priorities. By prioritizing sockets in this manner, background communication tasks such as automatic software upgrades may be conducted by the service 100 transparently to the end user.

[0029] Three embodiments of methods for prioritizing data transports are illustrated in FIGS. 4a, 4b and 4c. These methods may be implemented by the specific system architecture illustrated in FIG. 3, or by a different system architecture. FIGS. 4a and 4b illustrate methods in which sockets are prioritized as either “background” sockets or “foreground” sockets. As used herein, “foreground” sockets are sockets generated for interactive applications (e.g., Web browsing, instant messaging, . . . etc) whereas “background” sockets are sockets generated for non-interactive (or less interactive) applications (e.g., software upgrades).

[0030] Referring to FIG. 4a, at 405, a new background socket connection is initiated for a non-interactive application. For example, a specified portion of the software or firmware on the data processing device may need to be upgraded by the service 100. In this case, a background socket connection (e.g., sockets 311, 313) may be established between an upgrade application on the device 110 (e.g., application 391) and an upgrade service module on the service 100 (e.g., service module 392).

[0031] At 410, the socket prioritization logic 330, 331 on the device 110 and/or the service 100, respectively, determines whether any foreground sockets are currently supporting interactive applications (e.g., application 390). If not, then at 430, a background socket is generated for the non-interactive application with a specified maximum window size (‘m’). In one embodiment, the maximum window size may be set to the maximum size supported by the transport layer 305, 306 (e.g., 64 KB for standard TCP). In another embodiment, however, the maximum window size may be limited to a smaller value (e.g., 8 KB) so that, if/when the user begins performing interactive functions on the data processing device 110, the device can respond without an unreasonable delay (i.e., which might result if the device were required to wait for a maximum-sized window to be received before initiating the foreground socket).

[0032] At 410, if the socket prioritization logic 330, 331 determines that foreground sockets exist, then at 420, the background socket may be generated with a relatively smaller maximum window size (‘n’). In one embodiment, the maximum window size n is set to zero. However, in another embodiment, the maximum window size may be set to a relatively small value (e.g., 2 KB).

[0033] Moreover, in one embodiment, in order to allocate a specific percentage of the overall bandwidth to the background socket, the maximum window size for that socket may be reset to zero periodically. For example, if the total bandwidth available to the data processing device is 32 KB/sec, to allocate ¼ of the total bandwidth to a background socket, the socket prioritization logic 330, 331 may set the maximum window size to 8K once each second. When the 8K has been transmitted successfully (presumably in less than 1 second), the socket prioritization logic 330, 331 resets the maximum window size to zero until the beginning of the next second. This technique of precisely allocating a percentage of the overall bandwidth to a socket is illustrated graphically in FIG. 5. This bandwidth allocation technique may be employed at the various stages described herein where the maximum window size of a socket is set to a particular value (e.g., at 420 in FIG. 4a).

[0034] The socket prioritization logic 330, 331 may monitor the foreground socket connections to determine whether the user is still working interactively on the device 110. For example, in one embodiment, a foreground inactivity timer is reset to a threshold value (e.g., 30 seconds) each time the foreground socket is utilized (e.g., each time the user clicks on a new URL to a Web page). If the foreground socket is not utilized for the threshold period of time, determined at 425, then the maximum window size of the background socket may be increased (e.g., and reset to zero periodically based on the percentage of bandwidth to be allocated, as described above).

[0035] In the embodiment illustrated in FIG. 4a, the maximum window size is increased to ‘m’ if the inactivity threshold is reached (i.e., the value to which the maximum window size is set if no foreground sockets are initially detected). It should be noted, however, that the maximum window size may be increased to alternate values while still complying with the underlying principles of the invention. In one embodiment, the maximum window size may be increased incrementally over time, under the assumption that the longer the foreground socket is inactive, the less likely it is that the foreground socket will become active at any given point in time.

[0036] FIG. 4b illustrates a similar method for initiating a new foreground socket. At 440, the new foreground socket is initiated. If no background sockets exist, determined at 445, then at 447 the foreground socket is generated using a specified window size (e.g., 32 KB, 64 KB, . . . etc). If, however, background sockets exist, then the socket prioritization logic 330, 331 may reduce the bandwidth allocated to those sockets to free up bandwidth for the new foreground socket. Thus, at 450, the maximum window size for the background socket is set to ‘n’ and periodically reset to zero (as described above) to free up bandwidth. Once again, if the foreground socket becomes inactive, determined at 455, then the background socket window size may be reset to a larger value at 460 so that the background application can consume the additional, unused bandwidth.

[0037] In one embodiment of the invention, sockets are not strictly classified as “background” or “foreground” sockets as described above. Rather, sockets may be assigned a relative priority value (e.g., between 1 and 4) based on the level of interactivity of the applications which they support (or based on other specified variables).

[0038] One embodiment of a method for prioritizing socket connections in this manner is illustrated in FIG. 4c. At 470 a new socket connection is initiated with a specified priority (‘x’). At 475, the socket prioritization logic 330, 331 determines whether any other active sockets exist. If not, then the new socket is initiated in a typical manner, using a specified window size at 477.

[0039] If, however, active socket connections exist between the device 110 and the service 100 (or a different client/server) then, at 480, the priorities of all of the socket connections are evaluated and, at 485, the maximum window sizes are reallocated based on the relative priorities of the active sockets. For example, if three sockets exist having relative priorities of 1, 2 and 3 (with 1 being the highest priority and 3 being the lowest) then the socket with a priority of 1 may be allocated a larger window size than the socket with a priority of 2, and the socket with a priority of 2 may be allocated a larger window size than the socket with a priority of 4. If the maximum available bandwidth is 64 KB/sec, then the highest priority (1) socket may be allocated, for example, a 40 KB window, the lowest priority (3) socket may be allocated an 8 KB window and the socket with an intermediate priority (2) may be allocated a 16 KB window. If each socket is reset to zero each second as described above, then the 64 KB/sec bandwidth will be accurately divided among the sockets based on relative priority. In one embodiment, however, the highest priority socket may not be periodically reset to zero as described, so that it will be provided with the maximum available bandwidth.

[0040] Once again, if a particular socket becomes inactive for a predetermined period of time (e.g., because a user stops interactively working on an application), determined at 490, then, at 495, the window sizes for the remaining active sockets may be reallocated based on their relative priorities (as described above).

[0041] Embodiments of the invention may include various steps as set forth above. The steps may be embodied in machine-executable instructions which cause a general-purpose or special-purpose processor to perform certain steps. Alternatively, these steps may be performed by specific hardware components that contain hardwired logic for performing the steps, or by any combination of programmed computer components and custom hardware components.

[0042] Elements of the present invention may also be provided as a machine-readable medium for storing the machine-executable instructions. The machine-readable medium may include, but is not limited to, floppy diskettes, optical disks, CD-ROMs, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, magnetic or optical cards, propagation media or other type of media suitable for storing or transmitting electronic instructions. For example, the present invention may be downloaded as a computer program which may be transferred from a remote computer (e.g., a server) to a requesting computer (e.g., a client) by way of data signals embodied in a carrier wave or other propagation medium via a communication link (e.g., a modem or network connection).

[0043] Throughout the foregoing description, for the purposes of explanation, numerous specific details were set forth in order to provide a thorough understanding of the invention. It will be apparent, however, to one skilled in the art that the invention may be practiced without some of these specific details. For example, while embodiments of the invention described above operate at the TCP transport layer, the underlying principles of the invention are not limited to any particular data transport or any particular network layer. Moreover, while described above in the context of a data processing device 110 communicating with a data service 100, the underlying principles of the invention may be employed between virtually any two nodes communicating over a data network. Accordingly, the scope and spirit of the invention should be judged in terms of the claims which follow.