Browse Prior Art Database

Method for Dynamically Routing Web Requests to Different Web Servers

IP.com Disclosure Number: IPCOM000119118D
Original Publication Date: 1997-Dec-01
Included in the Prior Art Database: 2005-Apr-01
Document File: 4 page(s) / 174K

Publishing Venue

IBM

Related People

Iyengar, AK: AUTHOR

Abstract

Disclosed is a method for routing Web requests to specific Web servers in a cluster containing multiple Web servers which allows the Web server to which a request is directed to be determined dynamically. This method is especially useful when different Web servers contain different information. In this situation, dynamic routing is needed in order to direct requests to Web servers with the proper content to satisfy a request.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 30% of the total text.

Method for Dynamically Routing Web Requests to Different Web Servers

      Disclosed is a method for routing Web requests to specific Web
servers in a cluster containing multiple Web servers which allows the
Web server to which a request is directed to be determined
dynamically.  This method is especially useful when different Web
servers contain different information.  In this situation, dynamic
routing is needed in order to direct requests to Web servers with the
proper content to satisfy a request.

There are two components to the methodology:
  1.  Node-specific routing by making changes to the
       application.  No changes to the server are required.
  2.  Node-specific routing by making changes to the server.  No
       changes to the application are required.

It is possible to use one component without the other.  Ideally, both
components should be used.

      Node-specific routing by the application

      The server application generates hypertext links which clients
select to continue conversations.  A conversation is a series of
interactions between a client and one or more servers in which the
client selects the next page to view by following a hypertext link
provided by a server.  This method applies to hypertext links for
which the correct server is not known at the time the application is
written.  Such hypertext links are referred to as "partially
unknown".  Partially unknown hypertext links are identified by the
comment line "partially_unknown" immediately preceding the hypertext
link.  The application assigns a server number to partially unknown
hypertext links based on their pathnames.  The idea is that all
partially unknown hypertext links which have the same number will
point to the same server which is unknown when the application is
written.  Each partially unknown hypertext link will have a default
server name to which the request will be made if the server number
corresponding to the link has not yet been assigned to a server.  For
example,
       <!--partially_unknown--!><a     href="http://
       tranman.location.company.com/skyline/product-info.html">
  represents  a  partially  unknown hypertext link with a server
number determined by the pathname:
       //tranman.location.company.com/skyline/product-info.html.

The 'tranman.location.company.com' is the default server for this
hypertext link.  If the server corresponding to the server number is
unknown at the time the server sends the link to the client, no
changes are made to the link and the request is directed to
tranman.location.company.com.  Otherwise, the hypertext link is
modified to point to the server assigned to the server number.

      When  the  application is built, a table linking pathnames with
server numbers is created.  This table is known as the server number
table.

      A  key enabler for our algorithm is a dynamic argument embedder
which preserves state vari...