User configurable prefetch control system for enabling client to prefetch documents from a network server

Abstract

A prefetching and control system for a computer network environment. The user configures the client's prefetch parameters which are: enabling/disabling prefetching, prefetch threshold value, and the maximum number of documents to prefetch. A prefetch value or weight is contained in the Hypertext Markup Language (HTML) page or prefetch file, called a pathfile, for each link. The HTML page contains the prefetch values for each of its links, while pathfile contains the weights for every link on the HTML page associated with the Universal Resource Locator (URL). The client compares the prefetch or weight values of each link with its threshold value to decide if the link should be prefetched and placed in the local cache as long as the maximum number of documents to prefetch is not exceeded. Pathfiles reside on the server and are created by the server or web administrator/author. The server automatically creates the pathfiles from its log files which are filtered to retain all of the valid document requests and average paths are derived from the filtered results. Weights are assigned to each path in the URL by the server and inserted into the pathfile along with the associated paths. If no log files exist on the server, then the web administrator/author may manually enter in the weights for each path.

Claims

I claim:

1. A process for prefetching documents in a computer network environment, comprising the steps of:

configuring prefetch parameters of a client;

parsing a document for prefetch values;

comparing said prefetch parameters with said prefetch values; and

prefetching documents from a server onto said client;

wherein said prefetch values determine the relative weight of each link in said document.

2. The process of claim 1, wherein said prefetch parameters include enabling and disabling the prefetch.

3. The process of claim 1, wherein said prefetch parameters include a prefetch threshold value.

4. The process of claim 1, wherein said prefetch parameters include a document prefetch limit, limiting the amount of documents that said client prefetch.

5. The process of claim 1, said parsing step further comprising the steps of:

parsing the Hypertext Markup Language (HTML) document for prefetch values;

requesting a prefetch file from said server if no prefetch values exist in said HTML document and said prefetch file exists on said server and said prefetch file does not exist in the local cache;

receiving said prefetch file on said client;

placing said prefetch file into said local cache; and

parsing said prefetch file for prefetch values;

wherein said HTML document contains prefetch values for each document link in said HTML document; and

wherein said prefetch file contains paths within the HTML page associated with the Universal Resource Locator (URL) and their associated prefetch values.

6. The process of claim 1, said comparing step further comprising the step of:

selecting a link with said prefetch values greater than said prefetch parameters.

7. The process of claim 4, said prefetch step further comprising the steps of:

retrieving a link from said server if said link does not exist in the local cache;

storing said link in said local cache; and

stopping said prefetch process when the number of prefetched documents has reached said document prefetch limit or no more links exist;

wherein said prefetch process is repeated for each qualified link.

8. The process of claim 1, further comprising the step of:

assigning a greater weight to any links in the vicinity of the user's mouse position on the computer screen.

9. The process of claim 1, further comprising the steps of:

querying each link in said document for keywords; and

assigning a greater weight to links containing a common keyword.

10. The process of claim 1, further comprising the step of:

assigning a greater weight to links contained in a sequential pattern.

11. The process of claim 1, further comprising the step of:

assigning greater weights to links contained in a hub and spoke pattern.

12. The process of claim 5, wherein said server notifies said client if said prefetch file exists.

13. The process of claim 5, wherein said server includes said prefetch file in a multi-part message.

14. The process of claim 1, wherein said client notifies said server if said client is configured for prefetching.

15. An apparatus for prefetching documents in a computer network environment, comprising:

a module for configuring the prefetch parameters of a client;

a module for parsing a document for prefetch values; a module for comparing said prefetch parameters with said prefetch values; and

a module for prefetching documents from a server onto said client;

wherein said prefetch values determine the relative weight of each link in said document.

16. The apparatus of claim 15, wherein said prefetch parameters include enabling and disabling the prefetch.

17. The apparatus of claim 15, wherein said prefetch parameters include a prefetch threshold value.

18. The apparatus of claim 15, wherein said prefetch parameters include a document prefetch limit, limiting the amount of documents that said client prefetch.

19. The apparatus of claim 15, said parsing module further comprising:

a module for parsing the Hypertext Markup Language (HTML) document for prefetch values;

a module for requesting a prefetch file from said server if no prefetch values exist in said HTML document and said prefetch file exists on said server and said prefetch file does not exist in the local cache;

a module for receiving said prefetch file on said client;

a module for placing said prefetch file into said local cache; and

a module for parsing said prefetch file for prefetch values;

wherein said HTML document contains prefetch values for each document link in said HTML document; and

wherein said prefetch file contains paths within a Universal Resource Locator (URL) and their associated prefetch values.

20. The apparatus of claim 15, said comparing module further comprising:

a module for selecting a link with said prefetch values greater than said prefetch parameters.

21. The apparatus of claim 18, said prefetch module further comprising:

a module for retrieving a link from said server if said link does not exist in the local cache;

a module for storing said link in said local cache; and

a module for stopping said prefetch module when the number of prefetched documents has reached said document prefetch limit or no more links exist;

wherein said prefetch module is repeated for each qualified link.

22. The apparatus of claim 15, further comprising:

a module for assigning a greater weight to any links in the vicinity of the user's mouse position on the computer screen.

23. The apparatus of claim 15, further comprising:

a module for querying each link in said document for keywords; and

a module for assigning a greater weight to links containing a common keyword.

24. The apparatus of claim 15, further comprising:

a module for assigning a greater weight to links contained in a sequential pattern.

25. The apparatus of claim 15, further comprising:

a module for assigning greater weights to links contained in a hub and spoke pattern.

26. The apparatus of claim 19, wherein said server notifies said client if said prefetch file exists.

27. The apparatus of claim 19, wherein said server includes said prefetch file in a multi-part message.

28. The apparatus of claim 15, wherein said client notifies said server if said client is configured for prefetching.

Description

BACKGROUND OF THE INVENTION

1. Technical Field

The invention relates to the performance enhancement of network browsers in a computer environment. More particularly, the invention relates to the intelligent prefetching of documents across a computer network, efficiently using the free bandwidth and enhancing perceived performance on the client.

2. Description of the Prior Art

The popularity of the Internet has exploded in recent years. Typical users are based in their homes with Internet access through modems. Unfortunately, modem access speeds are relatively slow, causing Internet browsers to respond sluggishly. Slow servers and networks also contribute to this latency. The latency on the client is measured by the time lag that is perceived by the user. It is the time period from the moment that the user first clicks on a document link to the point in time when the document actually appears on his screen.

When the user clicks on a particular link, the Internet browser pulls a web document off of a web site. The browser does this by instantiating a connection with the web site server and pulls the document down from the server.

One approach to improving the perceived performance of the browser is by prefetching documents. Prefetching is the retrieval of documents from the web before the user makes an explicit request for them. The links are pulled down from the server onto the client and placed in a cache on the client. When the user actually clicks on a link, it comes from the cache rather than the network. The perceived performance is increased because the user thinks that the document is coming off the network when it actually is coming from his cache.

Many Internet browsers and add-on programs currently perform a "blind" prefetch, whereupon the browser pulls down all of the document links on a web page before the user has actually clicked on them. For example, in a very simple model of blind prefetching, the client pulls down page A from a server, parses the document links on that page and pulls each of them into the local cache. When the user clicks on a link to page B, it is loaded from the prefetched copy in the cache.

One of the problems with the blind prefetch approach is that it is very inefficient. The client must make multiple requests for all of the document links on a page. The server must respond to each request by sending the appropriate document, thereby forcing the server to send many more documents than are needed. This scenario is repeated for every client communicating with the server. The result is high network traffic and overloading of the server. Further, the client has no control over what is being pulled down from the server in this approach; it blindly makes the requests.

It would be advantageous to provide a prefetching and control system that improves the user-perceived network performance by allowing the client to efficiently decide what documents to prefetch, thereby reducing the demand on the server and network. It would further be advantageous to provide a prefetching and control system that enables the server to designate to the client what document links are important or the most common on a page.

SUMMARY OF THE INVENTION

The invention provides a prefetching and control system. The invention uses an easily configured prefetching system that allows a client to control the amount of documents that are prefetched from a server in a computer network environment. In addition, the invention uses a server-based control system that enables the server to pre-define the priority of each document link on a web page and notify a client of those priorities.

In a preferred embodiment of the invention, the user configures the client's prefetch parameters which are: enabling/disabling prefetching, prefetch threshold value, and the maximum number of documents to prefetch. A prefetch value or weight is contained in the Hypertext Markup Language (HTML) page or prefetch file, called a pathfile, for each link. The HTML page contains the prefetch values for each of its links, while pathfile contains the weights for every link on the HTML page associated with the Universal Resource Locator (URL). The client compares the prefetch or weight values of each link with its threshold value to decide if the link should be prefetched. If the value of the link is greater than the threshold value, then the link is prefetched and placed in the local cache as long as the maximum number of documents to prefetch is not exceeded. The client efficiently uses its free bandwidth by prefetching the documents that are most likely to be read by the user, thereby improving the user-perceived network performance and reducing the demand on the server and network.

Pathfiles reside on the server and are created by the server or web administrator/author. The server automatically creates the pathfiles from its log files. The log files are filtered to retain all of the valid document requests and average paths are derived from the filtered results. Weights are assigned to each path in the URL by the server and inserted into the pathfile along with the associated paths. If no log files exist on the server, then the web administrator/author may manually enter in the weights for each path. Pathfiles give the server the power to control client access to its resources.

Other aspects and advantages of the invention will become apparent from the following detailed description in combination with the accompanying drawings, illustrating, by way of example, the principles of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of a web page containing document links according to the invention;

FIG. 2 is a block schematic diagram of a document request/reply sequence of an Internet based client/server configuration according to the invention;

FIG. 3 is a block schematic diagram of the contents of a document request and a document packet according to the invention;

FIG. 4 is a block schematic diagram of a client prefetch sequence according to the invention;

FIG. 5 is a block schematic diagram of a client prefetch sequence using an Hypertext Markup Language (HTML) document according to the invention;

FIG. 6 is a block schematic diagram of a client prefetch sequence using a pathfile according to the invention; and

FIG. 7 is a block schematic diagram of a server pathfile creation sequence according to the invention.

DETAILED DESCRIPTION OF THE INVENTION

As shown in the drawings for purposes of illustration, the invention is embodied in a prefetch and control system for Internet client/server applications. A system according to the invention provides a configurable document prefetch in both the client and server in a computer network, thereby enhancing the user-perceived network performance. Many existing Internet browsers and add-on programs perform a "blind" document prefetch and provide little or no client or server control.

Referring to FIG. 1, a browser retrieves a page 101. During a blind prefetch, the browser requests the documents for all of the links on the page 101. Here, the browser requests the documents for the Developers 102, Learning Site 103, Product Training 104, AppFoundry 105, Tech Support 106, and Download 107 links from the server and places them in the local cache.

A prefetch and control system according to the invention includes a user-configurable prefetch limit on the client, combined with a prefetch file, called a pathfile, and a Hypertext Markup Language (HTML) web page document. The pathfile and HTML page are used by the client to decide what documents to prefetch. The server or web administrator/author creates the pathfile which the client retrieves from the server. The HTML page is created by the web author. This combination of configurability and pathfile/HTML page control efficiently uses the free network bandwidth on the client and lessens the load on the server, while additionally allowing the server to limit client access to its resources.

A user typically requests a page from a website at least once during normal browser use. Documents are prefetched while the user reads the page. The free bandwidth of the client is taken advantage of by prefetching to the local cache. A separate request for each prefetched document must be made by the client to the server. The invention allows the user to configure the client to prefetch a limited number of documents, which reduces the cache memory demand and lessens the load on the server through a smaller number of document requests. The user also configures the threshold value, which is the value that the client uses to decide which documents to prefetch.

With respect to FIGS. 2 and 3, the client 201 sends a document request 202 to the server 203. The document request packet 301 contains a document request 302 (the GET command) along with information, such as the client's capabilities, i.e. what the client is capable of handling (for example, what MIME types 303). The server 203 responds with the requested document 204. The HTML document 306 has an Hypertext Transfer Protocol (HTTP) header 304 attached. In this example, within the header 304 is an extension to the LINK tag 305 that tells the client 201 that if the client 201 is prefetch enabled, then it can find the pathfile for this Universal Resource Locator (URL) at "http://www.netscape.com/pathfilestxy" 309. The client 201 requests that document 309 from the server 203 if it has been enabled for prefetching and is unable to parse the prefetch values 308 in the HTML document 306.

The client 201 uses the prefetch values 308 in the HTML document 306 or pathfile 309 to decide which links on the page are the most likely to be accessed by a typical user. It prefetches the number of documents that it has been configured for from the server 203 according to the threshold value that the user configured on the client 201. Documents are prefetched that have prefetch values above the user-configured threshold value. The HTML document 306 contains prefetch values 308 for each document link contained on that page, while the pathfile 309 contains the prefetch values for every link in the URL.

A prefetching client first attempts to parse the contents and extract the prefetch information from the ANCHOR and the LINK (described later in this text) tags. The ANCHOR tag appears in the HTML document.

The following is an example of a specification for the ANCHOR tag:

Extension to HTML:

The client prefetching involves the following enhancement to the anchor <A> tag:

    ______________________________________
    Subtag         Description
    ______________________________________
    NAME           Named link end.
    HREF                URL for link resource.
    REL                  Relationship forward link type.
    REV                  Reverse link.
    TITLE              Advisory title string
    PRE                  Prefetch value.
    ______________________________________

    ______________________________________
           Realm:<CR.vertline.CRLF.vertline.LF>
           /[<relative path name>]<CR.vertline.CRLF.vertline.LF>
    ______________________________________

    ______________________________________
    filename:<CR.vertline.CRLF.vertline.LF>
    <prefetch document URL>[<space><weight>]<CR.vertline.CRLF.vertline.LF>
    [<prefetch document URL>[<space><weight<]CR.vertline.CRLF.vertline.LF>]
    <CR.vertline.CRLF.vertline.LF>
    ______________________________________

    ______________________________________
    # Pathfile for http://dummy.server.com/
    # Generated manually by Gagan Saksena on Aug. 1st. '97
    # Other comments here.
    # Realm has the relative path from the server name
    Realm:
    # The index.html file has links to several files.
    # It lists here all the relevant ones. Note that there
    # maybe other links/files not mentioned here.
    index.html:
    /whatsnew.html
    /help.html
    /contents.html
    /chap1/first.html
    whatsnew.html:
    /contents.html 0.7
    /help.html 0.3
    # Note implicit values - here 0.5
    contents.html:
    /index.html
    /chap1/first.html
    # Note a completely different file that may not even be
    # linked by any of the existing ones but has
    # links back to them.
    secret.html:
    /index.html
    /pics/family.html
    ______________________________________

    ______________________________________
    # Pathfile for http://gagan/
    # Generated manually by Gagan Saksena on Aug. 1st. '97
    # Other comments here.
    Realm:
    /chap1
    index.html:
    /index.html 0.2
    /contents.html 0.2
    first.html 0.6
    first.html:
    second.html 0.9
    /contents.html 0.1
    second.html:
    first.html
    /chap2/first.html
    ______________________________________

    ______________________________________
    if (prefetching.enabled)
      //Construct the corresponding pathfile URL from
    thisURL
      String pathfileLocation =
    extractPathfileLocationFromHeaders(thisURL);
      //If the constructed pathfileLocation is not in cache
      if (!cache.inCache(pathfileLocation))
      {
      //Download and add to cache
      pathfile = fetch(pathfileLocation);
      cache.add(pathfileLocation, pathtile);
      }
      //Parse the pathfile
      PathfileParser parser = new PathfileParser(pathfile);
      //If valid
      if (parser.contains((thisURL)))
      {
      //Construct the list to be prefetched for thisURL
      ListToPrefetch list = parser.list(thisURL,
    prefetchThreshold);
      //Retrieve all of them
      while (link=list.getNext())
      {
       //Add to cache
       cache.add(link, fetch(link));
      }
     }
    }
    ______________________________________

• Inventors
  Saksena, Gagan;
• Assignee
  Netscape Communications Corporation (Moutain View, CA)
• Published
  Feb-8-2000
• Current US Classes:
  709/219
  709/232
  719/310
• Application #
  009324
• International Classes
  G06F 013/00
• Field of Search
  709/201 709/203 709/217 709/218 709/219 709/230 709/231 709/232 709/234 709/250 709/300 709/302 711/137 711/204 77/10 77/104
• Examiner
  Vu; Viet D.
• Agent
  Glenn; Michael A.
• US Patent References:
  5758076
  5802292
  5805815
  5903727