US 6,243,757 B1Grant

Automated information filtering and distribution system

Issue Date:2001-06-05

•8 Claims

•13 Drawing Sheets

Abstract

An efficient and scalable real-time information distribution system and method of operation thereof are provided that build customized information collections according to individual preferences. This distribution system uses an information distribution network such as the Internet for its communications infrastructure. The system of the present invention is able to achieve high scalability and efficiency by routing information through the information distribution network. Information streams enter the information distribution network through feed processors. Feed processors, in turn, parse the information received from feed sources and create a set of “keyples.” The keyples are then passed on to a series of keyple routers. The keyple routers match the keyples to a set of destinations, thus multiplexing the keyples to only those destinations that have requested the information. A destination can consist of either another keyple router (which allows for the information distribution network to scale to a massive number of destinations) or a keyple customizer. A keyple customizer assigns incoming keyples to collections and then passes them on to a collection builder. A collection builder, in turn, constructs custom keyple collections for individual users.

Metadata

Assignee

eNuntio, Inc.

Inventors

Rajendra Kumar Kanodia
Steven Todd Black

Application Information

Application Number:US 09/229,393

Filing Date:1999-01-11

Priority Date:1999-01-11

Art Unit:7

Classifications

IPC:

G06F 1516

Field of Search:

709203709217709219709218709225709226709229709235

Patent Drawings (13 sheets)

Description

Cross-Reference to Appendix A

[0002] Appendix A, which is part of the present disclosure, contains a listing of segments of a computer program and related data, according to an embodiment of the invention. These listings of computer program contain 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 present disclosure, as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all copyright rights whatsoever.

Background of the Invention

[0003] 1. Field of the Invention

[0004] The present invention relates generally to information processing systems and, in particular, to information filtering and distribution systems.

[0005] 2. Related Art

[0006] The Internet provides a forum for the distribution of information on a global basis. However, typically users are required to actively seek information either by specifying a given website or by performing a search of the information available on the world wide web. This model presumes that the user is already aware of the existence of the information. This model, therefore, does not provide a means for automatically distributing information to interested parties. To solve this problem several “push” technologies have been proposed by companies such as PointCast, BackWeb, and Marimba. These solutions store information in databases on servers, which are polled and queried for new information by the clients at periodic intervals. This architecture, however, has certain inherent disadvantages. First, the database-centric nature of the architecture limits system scalability because the database must repeatedly process user queries. Second, only limited customization of the information delivered to each user is provided since user queries must be kept brief and simple. Third, repeated polling results in a waste of communication bandwidth. Finally, because of delays introduced by the polling interval, these systems are not suitable for the delivery of time-sensitive information.

[0007] Email distribution has also been used to address this problem. However, while allowing for personalized information distribution, e-mail distribution of information does not provide fast distribution capability for time-sensitive delivery and dramatically increases the cost of distributing information over the Internet.

[0008] There is thus a need for a system for distributing information over the Internet that allows for both a high level of customization in the information distributed and timely delivery of time-sensitive information, without significantly impacting the cost of distributing information over the Internet.

Summary of the Invention

[0009] The system and method of the present invention provide an efficient and scalable real-time information distribution system that builds customized information collections according to individual preferences. This distribution system uses an information distribution network such as the Internet for its communications infrastructure.

[0010] The system of the present invention is able to achieve high scalability and efficiency by routing information through the information distribution network. Information streams enter the information distribution network through feed processors. Feed processors, in turn, parse the information received from feed sources and create a set of “keyples.” The keyples are then passed on to a series of keyple routers. The keyple routers match the keyples to a set of destinations, thus multiplexing the keyples to only those destinations that have requested the information. A destination can consist of either another keyple router (which allows for the information distribution network to scale to a massive number of destinations) or a keyple customizer. A keyple customizer assigns incoming keyples to collections and then passes them on to a collection builder. A collection builder, in turn, constructs custom keyple collections for individual users. Finally, users can connect to collection servers and examine their collections, which contain only the information requested by the users.

Brief Description of the Drawings

[0011] FIG. 1 is a block diagram of a system for filtering and distributing information, in accordance to an embodiment of the invention.

[0012] FIG. 2A is a block diagram of a keyple, in accordance to an embodiment of the invention.

[0013] FIG. 2B is a block diagram of a customized keyple, in accordance to an embodiment of the invention.

[0014] FIG. 2C is a flow diagram of a keyple creation operation, in accordance to an embodiment of the invention.

[0015] FIG. 3A is a flow diagram of the operation of a feed processor of FIG. 1.

[0016] FIG. 3B is a data flow diagram of the operation of a feed processor of FIG. 1.

[0017] FIG. 4A is a flow diagram of the operation of a keyple router of FIG. 1.

[0018] FIG. 4B is a data flow diagram of the operation of a keyple router of FIG. 1.

[0019] FIG. 5A is a flow diagram of the operation of a keyple customizer within a collection processor of FIG. 1.

[0020] FIG. 5B is a data flow diagram of the operation of a keyple customizer within a collection processor of FIG. 1.

[0021] FIG. 6A is a flow diagram of the operation of a collection builder within a collection processor of FIG. 1.

[0022] FIG. 6B is a data flow diagram of the operation of a collection builder within a collection processor of FIG. 1.

[0023] FIG. 7A is a diagram of a web browser window displayed on a screen of a terminal computer of FIG. 1.

[0024] FIG. 7B is a diagram of a web browser window displayed on a screen of a terminal computer of FIG. 1 in response to a user selecting a hypertext link.

[0025] FIG. 8 shows a web browser window displayed on a screen of terminal computer of FIG. 1 to enable a user of the terminal computer to specify customization options.

Detailed Description of the Invention

[0026] FIG. 1 is a block diagram of a system 100 for filtering and distributing information, in accordance to the present invention. System 100 includes feed processors 110n(where n=A, B, C, etc.) that receive messages from feed sources 120n(where n=A, B, C, etc.). Messages received from feed sources 120nmay contain an abstract of the content represented by the message, a URL pointing back to a location in which the content is stored and one or more keywords (keys) describing the content. Feed processors 110n,in turn, either extract keys from the messages received from feed sources 120nor generate keys on-the-fly, when no keys are supplied from feed sources 120n.Feed sources 120nare any source of messages known in the art, such as news agencies, websites, stock brokerage companies, information services, etc. Messages, in turn, are any kind of structured or unstructured information suitable for mass distribution such as news, stock quotations, images, sound bytes, multimedia segments, etc.

[0027] The messages received from feed sources 120nare either stored on feed sources 120nor transmitted and cached on remotely located computers. In some embodiments, messages are delivered via satellite, RF transmission, network broadcast or other efficient mass distribution channel to servers located at Internet Service Providers (ISPs) sites. In other embodiments, the messages are stored on servers at a central location together with feed processors 110n.

[0028] Feed processors 110n,then, package the keys into keyples 200 (FIG. 2A). Keyples 200 are data sets that store information about keys. Each keyple 200 includes a version number 210, a feed ID 220, a message ID 230, a key ID 240, an event count 250 and a timestamp 260. In some embodiments, version number 210, feed ID 220, message ID 230, key ID 240, event count 250 and timestamp 260 require 4 bytes each, for a total of 24 bytes for an entire keyple 200. Since keyples 200 are dramatically reduced in size with respect to the messages (typical size 2048 bytes) they represent, the bandwidth required to transmit keyples 200 is substantially less than the one required to transmit the corresponding messages. As a result, mass distribution of keyples 200, unlike prior art “push” technologies, can be effected inexpensively over information distribution network 130.

[0029] Keyples 200 are then transmitted from feed processors 110nto collection processors 140n(where n=A, B, C, etc.) over information distribution network 130. Information distribution network 130, in turn, is made up by a series of keyple routers 160n(where n=A, B, C, etc.) connected via a computer network. Each keyple router 160n,in turn, receives keyples 200 and compares feed IDs 220 and key IDs 240 of each keyple 200 to feed ID/key ID pairs stored in a keyple router table 470 (FIG. 4B) to determine whether to route the keyple 200 to a destination specified by keyple router table 470. Destinations specified by keyple router table 470 can be either collection processors 140nor additional keyple routers 160n.Therefore, keyple routers 160ncan be organized in a hierarchical structure to allow fast routing of massive amounts of keyples 200.

[0030] If a match is found in keyple router table 470 to a destination corresponding to a collection processor 140n,the keyple 200 is routed to collection processor 140n.Collection processors 140n,in turn, receive keyples 200 and organize keyples 200 in customized keyple collections. Customized keyple collections are constructed in accordance to a set of preferences specified by users of system 100. The structure of a customized keyple 270 is shown in FIG. 2B. Customized keyple 270 is identical to keyple 200, except that a collection ID 280 is added.

[0031] Terminal computers 150n(where n=A, B, C, etc.) are connected to collection processors 140neither directly or through an ISP server. In the latter case, terminal computers 150nare connected to the ISP servers either through network connections or through dial-up modem connections. Users of terminal computers 150ncan access the messages represented by their customized keyple collections by connecting to the collection processors 140nusing a web browser program or other information retrieval technique known in the art. When the user connects to collection processor 140n,collection processor 140nextracts messages associated with customized keyples 280 and transmits customized keyples 280 to terminal computers 150n,where customized keyples 280 are displayed in the web browser window. The user can then retrieve the content associated with the message by selecting a hypertext link embedded in the message.

[0032] The content associated with the message is either retrieved directly from feed servers 120nor from an information cache on ISP servers. For example, in some embodiments the message contains a headline and a URL pointing to a location of the complete story which represents the content associated with the message. When the user follows the link associated with the headline, the URL contained in the message is used to retrieve the content from a remote location. The content is then displayed to the user together with other information extracted from the message (e.g., timestamp, feed source, abstract, etc.). In some instances (e.g. stock quotes), however, no content is associated with the message and therefore, only the information extracted from the message is presented to the user.

[0033] Feed processors 110n,collection processors 140nand keyple routers 160 are any suitable computers known in the art for receiving feeds over a computer network and extracting information keys from the feeds and distributing keys over an information distribution network. In some embodiments, feed processors 110n, collection processors 140nand keyple routers 160 are any Pentium® or PentiumII® based computers, available from a variety of third party manufacturers, G3 or Macintosh® computers, available from Apple Computer, Inc. of Cupertino, Calif., or high-end workstations such as UltraSparc™ workstations, available from Sun Microsystems, Inc. of Mountain View, Calif.

[0034] Furthermore, keyple routers 160ncan be integrated with traditional network routers available from, for example, Cisco Systems, Inc. of San Jose, Calif. or 3COM Corporation of Santa Clara, Calif., that are part of information distribution network 130.

[0035] FIG. 2C is a flow diagram of a keyple creation operation 205 used to create keyple 200 on a feed processor 110n.In FIG. 2C, a block in a memory of feed processor 110nis first allocated to store keyple 200 in stage 215. In stage 225, version number 210 is inserted in the memory block allocated in stage 215. Version number 210 is generated in response to a change in the structure of keyple 200. In stage 235, feed ID 220 is inserted in the memory block allocated in stage 215. Feed ID 220 identifies the feed source 120nfrom which the message represented by keyple 200 is extracted. In stage 245, message ID 230 is inserted in the memory block allocated in stage 215. Message ID 230 is a unique identifier assigned by feed processor 110nto each message received by feed processor 120n.In stage 255, key ID 240 is inserted in the memory block allocated in stage 215. Key ID 240 is a unique ID obtained by searching key table 320 for each key extracted from the message received from feed source 120n.Key ID 240 need not be unique across keys extracted from different feed sources 120n.However, each combination of key ID 240 and feed ID 220 must be unique. Since multiple keys can be extracted from a single message received from feed source 120n,more than one keyple 200 may be generated for each message.

[0036] In stage 265, event count 250 is inserted in the memory block allocated in stage 215. Event count 250 is a numerical value assigned by feed processor 120nrepresenting the number of messages received from a particular feed source 120nthat match key ID 240 of keyple 200. In stage 275, timestamp 260 is inserted in the memory block allocated in stage 215. Time stamp 260 is a value representing the time at which the message received from feed source 120noriginated. Finally, keyple 200 is generated in stage 285. The order in which stages 225-275 of operation 205 are performed is purely illustrative in nature. In fact, stage 225-275 of operation 205 can be performed in any order, in accordance to the principles of the present invention.

[0037] FIG. 3A is a block diagram of a message processing operation 300 performed by a feed processor 110n.First, in stage 305, feed processor 110nreceives a new message from a feed source 120n.Each message received by feed processor 110nmay contain one or more keys associated with the message received from feed source 120n.Alternatively, a wildcard key may be used to match any message received from a particular feed source 120n.Stage 310 determines whether any of the keys received as part of the message are unprocessed, in which case operation 300 proceeds to stage 315. Otherwise, operation 300 proceeds to stage 330. Stage 315 determines whether a key match is found by comparing a key received from feed source 120nto the entries stored in key table 320. If a key match is found, operation 300 proceeds to stage 325. Otherwise, operation 300 reverts back to stage 310. If a key match is found, a keyple is generated in stage 325 and operation 300 reverts to stage 310. Thus, if multiple key matches are found for a message received from feed source 120n,multiple keyples 200 are generated. Stage 330, then determines whether any keyples were generated for the message received from feed source 120n,in which case operation 300 proceeds to stage 335. Otherwise, operation 300 proceeds to stage 340 and operation 300 terminates. Finally, keyples 200 generated in stage 325 are transmitted to a keyple router 160nover information distribution network 130 and operation 300 terminates.

[0038] FIG. 3B illustrates data flow 350 through feed processor 110n.A feed stream received from feed source 120nis received by feed listener 355. Feed listener 355 extracts a message from the feed stream, stores the message in message queue 360 and signals to feed processor 365 that a new message has been received. Feed processor 365, in turn, extracts a message from message queue 360 and compares any keys contained in the message to the entries stored in key table 320. If any key matches are found, as explained with respect to the flow diagram of FIG. 3A, feed processor 365 adds a keyple 200 to keyple queue 375 and signals feed transmitter 380 that keyple 200 has been added to keyple queue 375. Finally, feed transmitter 380 periodically packages keyples 200 into keyple packets and transmits the keyple packets over information distribution network 130. Keyple packet transmissions are triggered either by having a sufficient number of keyples 200 to fill a keyple packet in keyple queue 375 or by expiration of a preset period of time in between keyple packet transmissions. However, some keyples, known as priority keyples, are immediately transmitted regardless of whether the keyple packet is filled. Thus, priority keyples can be used to transmit time sensitive information to ensure real time distribution of information.

[0039] In some embodiments, feed listener 355, feed processors 365 and feed transmitter 380 are separate threads of a C++ program executed by a Linux® (Red Hat v. 5.2) operating system running on feed processor 110n.Appendix A contains code segments illustrating possible implementations of feed listener 355, feed processor 365 and feed transmitter 380. Those skilled in the art, however, realize that feed listener 355, feed processor 365 and feed transmitter 380 could be implemented in computer languages other than C++ and executed by operating systems other than Linux®. For instance, feed listener 355, feed processor 365 and feed transmitter 380 could be implemented as JAVA™ bytecodes executed by any multi-threading operating system, such as versions of Unix® other than Linux®, Mac®OS, available from Apple Computer, Inc. of Cupertino, Calif., or Windows NT® 4.0, available from Microsoft Corp. Of Redmond, Wash.

[0040] FIG. 4A is a flow diagram of operation 400 executed by a keyple router 160nof FIG. 1. Initially, a keyple 200 is received by keyple router 160nover information distribution network 130 in stage 405. Stage 410 then determines whether a routing match is found by comparing feed ID 220 and key ID 240 of keyple 200 to the feed ID/key ID pairs stored in keyple router table 470. If a routing match is found, operation 400 proceeds to stage 415. Otherwise, keyple router 160nnotifies the sender of keyple 200 that no routing match has been found in stage 435. Thus, keyple processing is completed in stage 440 and operation 400 terminates. If a routing match is found, a destination list is generated in stage 415. Thus, more than one destination can be specified for each routing match. Stage 420 determines whether a copy of keyple 200 has been sent to all destinations in the destination list generated in stage 415, in which case keyple processing is completed in stage 440 and operation 400 terminates. Otherwise, a copy of keyple 200 is generated in stage 425 and the copy of keyple 200 is transmitted to a new destination in stage 430. Stages 420-430 are then repeated until all destinations in the destination list have been processed. Keyple processing is completed in stage 440 and operation 400 terminates.

[0041] FIG. 4B illustrates data flow 450 through keyple router 160n.A keyple stream transmitted over information distribution network 130 is received by keyple router listener 455. Keyple router listener 455 extracts a keyple 200 from the keyple stream, stores keyple 200 in keyple queue 460 and signals to keyple router processor 465 that a keyple 200 has been received. Keyple router processor 465, in turn, extracts a keyple 200 from keyple queue 460 and compares feed ID 220 and key ID 240 contained in keyple 200 to the entries stored in keyple router table 470. If any routing matches are found, as explained with respect to the flow diagram of FIG. 4A, keyple router processor 465 adds keyple 200 to keyple queue 475 and signals keyple router transmitter 480 that keyple 200 has been added to keyple queue 475. Finally, keyple router transmitter 480 periodically packages keyples 200 into keyple packets and transmits the keyple packets over information distribution network 130. Keyple packet transmissions are triggered either by having a sufficient number of keyples 200 directed to a same destination to fill a keyple packet in keyple queue 475 or by expiration of a preset period of time in between keyple packet transmissions to a destination. As explained with respect to FIG. 3B, priority keyples are transmitted immediately regardless of whether the keyple packet is full.

[0042] In some embodiments, keyple router listener 455, keyple router processor 465 and keyple router transmitter 480 are separate threads of a C++ program executed by a Linux® (Red Hat v. 5.2) operating system running on keyple router 160n.Appendix A contains code segments illustrating possible implementations of keyple router listener 455, keyple router processor 465 and keyple router transmitter 480. Those skilled in the art, however, realize that keyple router listener 455, keyple router processor 465 and keyple router transmitter 480 could be implemented in computer languages other than C++ and executed by operating systems other than Linux®. For instance, keyple router listener 455, keyple router processor 465 and keyple router transmitter 480 could be implemented as JAVA™ bytecodes executed by any multi-threading operating system, such as versions of Unix® other than Linux®, Mac®OS, available from Apple Computer, Inc. of Cupertino, Calif., or Windows NT® 4.0, available from Microsoft Corp. Of Redmond, Wash.

[0043] FIG. 5A is a flow diagram of operation 500 executed by a keyple customizer of collection processor 140nof FIG. 1. Initially, a keyple 200 is received by collection processor 140nover information distribution network 130 in stage 505. Stage 510 then determines whether a customization match is found by comparing feed ID 220 and key ID 240 of keyple 200 to the feed ID/key ID pairs stored in customization definition table 570. If a customization match is found, operation 500 proceeds to stage 515. Otherwise, collection processor 140nnotifies the sender of keyple 200 that no customization match has been found in stage 540. Thus, keyple processing is completed in stage 545 and operation 500 terminates. If a customization match is found, a keyple collection list is generated in stage 515. Thus, more than one keyple collection can be specified for each customization match. Stage 520 determines whether all keyple collections in the keyple collection list generated in stage 515 have been processed, in which case keyple processing is completed in stage 545 and operation 500 terminates. Otherwise, a copy of keyple 200 is generated in stage 525. A customized keyple 270 is then generated in stage 530 by inserting collection ID 280 into the copy of keyple 200 generated in stage 525. Customized keyple 270 is transmitted in stage 535. Stages 520-535 are then repeated until all keyple collections in the keyple collection list have been processed. Thus, keyple processing is completed in stage 545 and operation 500 terminates.

[0044] FIG. 5B illustrates data flow 550 through a keyple customizer of collection processor 140n.A keyple stream transmitted over information distribution network 130 is received by customizer listener 555. Customizer listener 555 extracts a keyple 200 from the keyple stream, stores keyple 200 in keyple queue 560 and signals to customizer processor 565 that a keyple 200 has been received. Customizer processor 565, in turn, extracts a keyple 200 from keyple queue 560 and compares feed ID 220 and key ID 240 contained in keyple 200 to the entries stored in customization definition table 570. If any customization matches are found, as explained with respect to the flow diagram of FIG. 5A, customizer processor 565 adds customized keyple 270 to customized keyple queue 575 and signals customizer transmitter 580 that customized keyple 270 has been added to customized keyple queue 575. Finally, customizer transmitter 580 sends customized keyple 270 to a collection builder of collection processor 140n.

[0045] In some embodiments, customizer listener 555, customizer processor 565 and customizer transmitter 580 are separate threads of a C++ program executed by a Linux® (Red Hat v. 5.2) operating system running on collection processor 140n.Appendix A contains code segments illustrating possible implementations of customizer listener 555, customizer processor 565 and customizer transmitter 580. Those skilled in the art, however, realize that customizer listener 555, customizer processor 565 and customizer transmitter 580 could be implemented in computer languages other than C++ and executed by operating systems other than Linux®. For instance, customizer listener 555, customizer processor 565 and customizer transmitter 580 could be implemented as JAVA™ bytecodes executed by any multi-threading operating system, such as versions of Unix® other than Linux®, Mac®OS, available from Apple Computer, Inc. of Cupertino, Calif., or Windows NT® 4.0, available from Microsoft Corp. Of Redmond, Wash.

[0046] FIG. 6A is a flow diagram of operation 600 of a collection builder of collection processor 140n.A customized keyple 270 transmitted from a keyple customizer of collection processor 140nis first received by the collection builder in stage 605. Customized keyple 270 is stored in keyple cache 615 in stage 610. Cache 615 stores the most recently received customized keyples 270 for fast access. Customized keyples 270 are periodically flushed from keyple cache 615 or written to disk for permanent storage. Stage 620 then determines whether a user of a terminal computer 150nis connected to collection processor 140n,in which case operation 600 proceeds to stage 625. Otherwise, keyple processing is completed in stage 640 and operation 600 terminates. A message associated with customized keyple 270 is then retrieved in stage 625. Information is extracted from the message retrieved in stage 625 and encoded for transmission to terminal computer 150nin stage 630. Depending on the type of connection between terminal computer 150nand collection server 140n the type of information extracted from the message and the formatting of the information may change. For instance, if the user of terminal computer 150nis accessing collection server 140nat an ISP site using a web browser over a dial-up connection, the information would be formatted as an HTML page. Alternatively, the information could also be formatted as JAVA™ bytecodes or other format supported by terminal computer 150n.Thus, keyple processing in completed in stage 640 and operation 600 terminates.

[0047] FIG. 6B illustrates data flow 650 through a collection builder of collection processor 140n.A customized keyple stream transmitted from a keyple customizer of collection processor 140nis received by builder listener 655. Builder listener 655 extracts a customized keyple 270 from the customized keyple stream, stores customized keyple 270 in customized keyple queue 660 and signals builder processor 665 that a customized keyple 270 has been received. Builder processor 665, in turn, extracts a customized keyple 270 from customized keyple queue 660, stores customized keyple 270 in keyple cache 615 and signals collection server 680 that customized keyple 270 has been added to keyple cache 615. Finally, collection server 680 processes customized keyple 270 as described with respect to the flow diagram of FIG. 6A.

[0048] In some embodiments, builder listener 655, builder processor 665 and collection processor 680 are separate threads of a C++ program executed by a Linux® (Red Hat v. 5.2) operating system running on collection processor 140n.Appendix A contains code segments illustrating possible implementations of builder listener 655 and builder processor 665. Those skilled in the art, however, realize that builder listener 655, builder processor 665 and collection processor 680 could be implemented in computer languages other than C++ and executed by operating systems other than Linux®. For instance, builder listener 655, builder processor 665 and collection processor 680 could be implemented as JAVA™ bytecodes executed by any multi-threading operating system, such as versions of Unix other than Linux®, Mac®OS, available from Apple Computer, Inc. of Cupertino, Calif., or Windows NT® 4.0, available from Microsoft Corp. Of Redmond, Wash.

[0049] FIG. 7A is a diagram of a web browser window 700 displayed on a screen of terminal computer 150n.. Web browser window 700 is a window displayed by any web browser program known in the art such as Internet Explorer 4.0, available from Microsoft Corporation of Redmond, Wash., or Netscape Navigator® 4.0, available from Netscape Communications, Corp. of Mountain View, Calif. Window 700 has a category selection menu 710 and multiple headlines 720n(where n=A, B, C, etc.). Each headline 720nhas a timestamp 730n corresponding to timestamp 260 of customized keyple 270, hypertext link 740nand feed source 750n. A user of terminal computer 150ncan retrieve the information associated with headline 720nby clicking on hypertext link 730n, as shown in FIG. 7B.

[0050] FIG. 7B is a diagram of a web browser window 760 displayed on a screen of terminal computer 150nin response to a user selecting hypertext link 740A of FIG. 7A. Web browser window 760 is a window displayed by any web browser program known in the art such as Internet Explorer 4.0, available from Microsoft Corporation of Redmond, Wash., or Netscape Navigator® 4.0, available from Netscape Communications, Corp. of Mountain View, Calif. Web browser window 760 includes a title 765, a feed source 770, a timestamp 780, an author 790 and a body 795.

[0051] FIG. 8 shows a web browser window 800 displayed on a screen of terminal computer 150nto enable a user of terminal computer 150nto specify customization options to be used by the keyple customizer of collection processor 140n.Web browser window 800 is a window displayed by any web browser program known in the art such as Internet Explorer 4.0, available from Microsoft Corporation of Redmond, Wash., or Netscape Navigator® 4.0, available from Netscape Communications, Corp. of Mountain View, Calif. Web browser 800 includes category tabs 810n(where n=A, B, C, etc.). Each tab 810nhas a feed source menu 820n,a subscribe button 830n,a get all content button 840n,a get sections button 850nand a sections list 860n.Windows NT is a registered trademark of Microsoft Corporation of Redmond, Wash. Pentium and Pentium II are registered trademarks of Intel Corporation of Santa Clara, Calif. Macintosh and MacOS are registered trademarks of Apple Computer, Inc. of Cupertino, Calif. UltraSparc and JAVA are trademarks of Sun Microsystems, Inc. of Mountain View, Calif. Unix is a registered trademark of American Telephone and Telegraph Company of New York, N.Y. Linux is a registered trademark of Linus Torvalds. Netscape Navigator is a registered trademarek of Netscape Communications, Corp. of Mountain View, Calif.

[0052] Embodiments described above illustrate but do not limit the invention. In particular, the invention is not limited to the particular hardware described herein. Those skilled in the art realize that alternative information delivery means can be employed in lieu of the ones described herein in accordance to the principles of the present invention. Furthermore, the invention is not limited to the distribution of any particular type of information. For example, the invention can be used to distribute news, advertisement, marketing materials, etc. Similarly, the invention is not limited to the particular web pages described herein. In fact, those skilled in the art realize that different web pages or even alternative information delivery techniques such as voice, paging, etc. may be used in lieu of the web pages of FIGS. 7A, 7B and 8. Other embodiments and variations are within the scope of the invention, as defined by the following claims.

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

[Inline image omitted. See PDF.]

Claims

We claim:

1. A system for filtering and distributing information, the system comprising:

a distribution network;

one or more feed processors connected to the information distribution network, wherein each feed processor receives messages from an external information source, extracts at least one key from each message and, for each key, generates a data set containing the key (keyple);

one or more collection processors connected to the information distribution network, wherein each collection processor receives keyples transmitted by one or more feed processors over the distribution network, and generates one or more keyple collections; and

one or more computers connected to the collection processors, wherein each computer receives messages associated with one or more keyple collections generated by the collection processors, the feed processors package multiple keyples into keyple packets and transmit keyple packets over the information distribution network; and the information distribution network further comprises one or more keyple routers connected over the information distribution network, wherein each keyple router receives keyple packets from either a feed processor or another keyple router, extracts keyples from a keyple packet, and routes each keyple either to another keyple router or to a collection processor.

2. The system of claim 1, wherein each keyple router further replicates each keyple and routes each replicated keyple either to another keyple router or to a collection processor.

3. The system of claim 1, wherein the keyples are routed by extracting a feed/key pair from each keyple and comparing the feed/key pair to information stored in a routing table of the keyple router.

4. A method for filtering and distributing information, the method comprising:

one or more feed processors connected to an information distribution network receiving messages from an external information source, extracting at least one key from each message and generating a keyple for each key, wherein the keyple is a data set containing the key;

one or more collection processors connected to the information distribution network receiving keyples transmitted by one or more feed processors over the distribution network, and generating one or more keyple collections;

one or more computers connected to the collection processors receiving messages associated with one or more keyple collections generated by the collection processors; and

the feed processors packaging multiple keyples into keyple packets and transmitting keyple packets over the information distribution network,

wherein the information distribution network further comprises one or more keyple routers connected over the information distribution network, the method further comprising each keyple router receiving keyple packets from either a feed processor or another keyple router, extracting keyples from a keyple packet, and routing each keyple either to another keyple router or to a collection processor.

5. The method of claim 4, further comprising each keyple router replicating each keyple and routing each replicated keyple either to another keyple router or to a collection processor.

6. The method of claim 4, wherein routing the keyples further comprises extracting a feed/key pair from each keyple and comparing the feed/key pair to information stored in a routing table of the keyple router.

7. A computer data signal embodied in a carrier wave transmitted from a feed processor to a collection processor over an information distribution network, after the feed processor has received a message from a feed source, the signal comprising a keyple, the keyple comprising:

a feed/key pair;

a message ID;

an event count; and

a time stamp,

wherein the information distribution network further comprises one or more keyple routers connected over the information distribution network, the keyple routers routing keyples from the feed processors to the collection processors by comparing the feed/key pair of each keyple to information stored in a routing table of the keyple router.

8. The signal of claim 7, wherein each keyple router further replicates each keyple and routes each replicated keyple either to another keyple router or to a collection processor.

Patent Citations (20)

Patent	Date	Inventor	Cited By
US4815030	1989-03-01	Cross et al.	Applicant
US5428778	1995-06-01	Brookes	Examiner
US5535382	1996-07-01	Ogawa	Applicant
US5537586	1996-07-01	Amram et al.	Applicant
US5606693	1997-02-01	Nilsen et al.	Applicant
US5634053	1997-05-01	Noble et al.	Applicant
US5657461	1997-08-01	Harkins et al.	Applicant
US5689642	1997-11-01	Harkins et al.	Applicant
US5740421	1998-04-01	Palmon	Applicant
US5740549	1998-04-01	Reilly et al.	Examiner
US5752246	1998-05-01	Rogers et al.	Examiner
US5754938	1998-05-01	Herz et al.	Examiner
US5768528	1998-06-01	Stumm	Applicant
US5790790	1998-08-01	Smith et al.	Examiner
US5855020	1998-12-01	Kirsch	Applicant
US5862325	1999-01-01	Reed et al.	Applicant
US5867667	1999-02-01	Butman et al.	Applicant
US5867799	1999-02-01	Lang et al.	Applicant
US6026369	2000-02-01	Capek	Examiner
US6047327	2000-04-01	Tso et al.	Examiner

Non-Patent Literature (14)

Malan, et al., “Salamander: A Push-based Distribution Substrate For Internet Applications”, USENIX Symposium on Internet Technologies and Systems, pp. 171-181, Dec. 8-11, 1997 (1997).Applicant
Yan, et al., “Sift—A Tool for Wide-Area Information Dessemination”, USENIX Technical Conference—Jan. 16-20, 1995, pp. 177-186 (1995).Applicant
Cooperstock, et al., “Why Use a Fishing Line When You Have a Net? An Adaptive Multicast Data Distibution Protocal”, 1996 USENIX Technical Conference—Jan. 22-26, 1996—San Diego, CA, pp. 343-352 (1996).Applicant
Manber, et al., “WebGlimpse—Combining Browsing and Searching”, USENIX Association 1997 Annual Technical Conference, pp. 195-206, (1997).Applicant
Manber, et al., “The Search Broker”, USENIX Association USENIX Symposium on Internet Technologies and Systems—Dec. 8-11, 1997, pp. 231-239 (1997).Applicant
Huang, et al., “A Bandwidth-Sensitive Update Scheduling Method for Internet Push”, 18th International Conference on Distributed Computing Systems, IEEE Computer Society Technical Committee on Distributed Processing, May 26-29, 1998, pp. 303-310 (1998).Applicant
Tomasic, et al., “Scaling Heterogeneous Databases and the Design of Disco”, Proceedings of the 16th International Conference of Distributed Computing Systems, May 27-30, 1996, IEEE Computer Society Press, pp. 449-457 (1996).Applicant
De Roure, et al., “A Distributed Hypermedia Link Service”, Proceedings Third International Workshop on Services in Distributed and Networked Enviroments Jun. 3-4, 1996, IEEE Computer Society Technical Committee on Distributed Processing, pp. 156-161 (1996).Applicant
Kretser, et al., “Methodologies for Distributed Information Retrieval”, Proceedings 18th International Conference on Distributed Computing Systems May 26-29, 1998, IEEE Computer Society Technical Committee on Distributed Processing, pp. 66-76 (1998).Applicant
Underwood, et al., “User-centered push for timely information delivery”, Computer Networks and ISDN Systems 30, No. 1-7, pp. 33-40 (Apr. 1998).Applicant
Ardö, “A regional distributed WWW search and indexing service—the Desire way”, Computer Networks and ISDN Systems, vol. 30, No. 1-7, pp. 173-183 (Apr. 1998).Applicant
Hartman, et al., “Index-based hyperlinks”, Computer Networks and ISDN Systems, vol. 29, No. 8-13, pp. 1129-1135 (Sep. 1997).Applicant
Malan, et al., “Comparison of Two Middleware Data Dissemination Services in a Wide-Area Distributed System”, Proceedings of the 17th International Conference on Distributed Computing Systems, IEEE Computer Society Press, pp. 411-419 (May 27-30, 1997).Applicant
Liu, et al., “A Dynamic Query Scheduling Framework for Distributed and Evolving Information Systems”, Proceedings of the 17th International Conference on Distributed Computing Systems, IEEE Computer Society Press, pp. 474-481 (May 27-30, 1997).Applicant

Abstract