Accelerating web access by predicting user action

• US 6,055,569 A
• Filed: 01/27/1998
• Issued: 04/25/2000
• Est. Priority Date: 01/27/1998
• Status: Expired due to Term

• Current Assignee: Go Ahead Software, Inc.
• Original Assignee: Go Ahead Software, Inc.
• Inventors: Gravestock, Peter Gerard, O'Brien, Michael David
• Primary Examiner(s): Maung, Zarni
• Assistant Examiner(s): Tran, Philip B.
• Application Number:
  US09/014,013
• Time in Patent Office: 819 Days
• Field of Search: 709/203, 709/217, 709/218, 709/223, 709/224, 709/219
• US Class Current: 709/223
• CPC Class Codes: G06F 16/9574 of access to content, e.g. by caching

Abstract

A smart browser working in conjunction with a HTTP server that selectively downloads WWW pages into the browser'"'"'s memory cache. The determination of which pages to download is a function of a probability weight assigned to each link on a Web page. By evaluating that weight to a predetermined browser criteria, only those pages most probably to be downloaded are stored in the browser'"'"'s memory cache. The download is done in the background while the browser user is viewing the current Web page on the monitor. This greatly enhances the speed with which the viewer can "cruise" the Web while at the same time conserving system resources by not requiring the system to download all the possible links.

Claims

1. A method to predict, in a computer network, what information is next to be downloaded from a server into a memory cache of a client computer comprising the steps of:

contacting a server as requested by a client computer;

retrieving a page of information associated with the server address requested, said page of information containing at least one link to other pages of information;

displaying the page of information on said client computer;

reading a key word containing a probability weight encoded in a tag denoting said link;

matching said probability weight with a predetermined criteria residing in the client computer; and

downloading the information associated with said link into the memory cache of said client computer if said probability weight meets said criteria;

wherein the access time required to view the next page selected is greatly reduced.

2. A method to predict, in a computer network, what information is next to be downloaded from a server into a browser'"'"'s memory cache of a client computer comprising the steps of:

contacting a server that uses the HTTP protocol as requested by a client computer;

retrieving a page of information wherein the information associated with the server address requested is formatted in HTML or XML;

displaying the page of information on said client computer wherein the client computer is enabled with a Web browser;

interpreting the probability weight of what page is to be selected next by the client user wherein the probability weight is contained in a HTML key word encoded within each page link, wherein each page link is in the form of an URL address, displayed on the page on the client computer;

matching the probability weight with a predetermined criteria residing in the client'"'"'s computer browser software; and

downloading the page associated with said link into the memory cache of said client computer if said probability weight meets said criteria;

wherein the access time required to view the next page selected is greatly reduced, if the page selected by the client user matches the pages already retrieved and waiting in said browser memory cache.

4. The method of claim 2 further comprising the steps of:

having the predetermined criteria be adjusted by a client user'"'"'s network administrator;

wherein the network administrator can adjust the amount of network resources used depending on the network conditions.

5. The method of claim 2 further comprising the steps of:

reading the pages already downloaded into the client user'"'"'s browser cache;

reading the key words containing the probability weight of the links contained on said pages;

evaluating the probability weights of said links; and

further downloading the pages those links point to, into the client user'"'"'s browser memory cache.

6. The method of claim 2 further comprising the steps of:

adjusting the probability weight of the link manually based on counting which links are actually chosen by the client user'"'"'s upon visiting the page;

recording said count; and

adjusting the probability weight of each link depending on results of the recording.

7. The method of claim 2 further comprising the steps of:

adjusting the probability weight of the link dynamically, based on counting which links are actually chosen by the client user'"'"'s upon visiting the page;

recording said count; and

adjusting the probability weight of each link at periodic intervals depending on results of the recording.

8. The method of claim 2 further comprising the steps of:

counting which links are actually chosen by a particular client user upon visiting the page;

storing said count into the client users memory in the form of a "cookie";

retrieving said "cookie" upon a follow-up visit of the Web site by said particular client user; and

adjusting the probability weight of each link depending on results of the "cookie" information.

3. The method of claim 3 further comprising the steps of:

having the predetermined criteria be adjustable by the client user;

wherein the client user can adjust the amount of network resources used depending on the network conditions.

9. A system that predicts, in a computer network, what information is to be next downloaded from a server, the system comprising:

a client user'"'"'s computer, a server computer, a networked link between said client computer and said server computer;

the client computer with the capability to obtain and display information stored on the server computer;

the server computer storing information in the form of pages which in turn contain links to other pages of information;

said links having a probability factor, encoded within, that said links will be downloaded;

said client user'"'"'s computer enabled to interpret said probability factor;

said client user'"'"'s computer also enabled to match said probability factor to a predetermined criteria residing in said client user'"'"'s computer; and

said client user'"'"'s computer then downloading said information prior to being selected by user, into the browser cache of said client user'"'"'s computer.

10. The system in claim 9 further comprising:

the client user'"'"'s computer enabled by a browser to be able to read HTML or XML formatted Web pages;

the server using HTTP protocol;

the links using a URL address to point to another Web page, site or object;

the probability assigned to each link to be in the form of a key word encoded into said link;

the browser enabled to interpret said probability factor and match said probability factor against a predetermined criteria.

11. A system that predicts, in a computer network, what information is likely to be downloaded next from the server, the system comprising:

a client user'"'"'s computer, a server computer, a networked link between said client computer and said server computer;

the client computer having the means for obtaining and displaying information stored on the server computer;

the server computer having the means for storing information in the form of pages which in turn have the means for containing links to other pages of information;

said links having the means for having an associated probability factor encoded within said links that they will be selected by a client computer user;

said client computer having the means for interpreting said probability factor;

said client computer also having the means for matching said probability factor to a predetermined criteria residing in said client computer; and

said client computer then having the means for downloading said information prior to being selected by the user, into a memory cache of said client computer.

12. The system in claim 11 further comprising:

the client user'"'"'s computer enabled by a browser means for obtaining and displaying information stored on the server computer;

the server computer enabled by a HTTP formatted means for storing pages of information which in turn have a HTML or XML formatted means for containing links to other pages of information;

said links having a HTML or XML formatted means for having an assigned probability factor encoded within said links that predict if said links would be selected by said client computer for downloading;

said client computer having the browser means for interpreting said probability factor;

said client computer also having the browser means for matching said probability factor to a predetermined criterion residing in said client computer.