Browse Prior Art Database

Method for High Speed Routing of Virtual Web Content

IP.com Disclosure Number: IPCOM000015263D
Original Publication Date: 2001-Oct-24
Included in the Prior Art Database: 2003-Jun-20
Document File: 6 page(s) / 86K

Publishing Venue

IBM

Abstract

A program is disclosed that extends a Web proxy server’s routing functionality. The program routes virtual directories to other Web servers, allowing Web content to be distributed across multiple physical machines, while presenting a uniform domain address for accessing the Web content. This program may be used to proxy browser requests to an intranet, or to the Internet. When a virtual directory is mapped to an intranet Web server, this is beneficial for distributing content over multiple physical machines. When a virtual directory is mapped to the Internet, users can browse specific Web sites exposed by the proxy through the virtual directories, as though those Web sites were an extension to the proxy's content. In the Figure below, the listed directories are mapped to a Web server that is not owned by the proxy administrator. In this example, the proxy server, acme.com, has 3 virtual directories: http://acme.com/toys http://toys.com http://acme.com/cars http://cars.com http://acme.com/plants http://plants.com

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 42% of the total text.

Page 1 of 6

Method for High Speed Routing of Virtual Web Content

A program is disclosed that extends a Web proxy
server's routing functionality. The program routes
virtual directories to other Web servers, allowing
Web content to be distributed across multiple
physical machines, while presenting a uniform domain
address for accessing the Web content.

This program may be used to proxy browser requests
to an intranet, or to the Internet. When a virtual
directory is mapped to an intranet Web server, this
is beneficial for distributing content over multiple
physical machines. When a virtual directory is
mapped to the Internet, users can browse specific
Web sites exposed by the proxy through the virtual
directories, as though those Web sites were an
extension to the proxy's content.

In the Figure below, the listed directories are
mapped to a Web server that is not owned by the
proxy administrator. In this example, the proxy
server, acme.com, has 3 virtual directories:

http://acme.com/toys -> http://toys.com
http://acme.com/cars -> http://cars.com
http://acme.com/plants -> http://plants.com

Web Browser

Web Server 1

toys.com

HTTP(S) Intranet HTTP(S)

Web Proxy Server acme.com

 /toys /cars /plants

Web Server 2

cars.com

Web Server 3

plants.com

Web Browser

Figure

The algorithm transparently maps the virtual

1

[This page contains 4 pictures or other non-text objects]

Page 2 of 6

directory to the destination Web server. No
configuration is required on the destination Web
server for this to work. Also, the routing is done
at high speed, because the content of the HTML data
sent to the browser is not rewritten. To the
browser, the request appears to be a local directory
on the proxy, while in reality, the content is
retrieved from another Web server.

Problem

Although the algorithm seems like a simple URL
mapping procedure, there are a number of complex
issues involved in solving this problem.

Typically, most of the content on Web pages have
absolute URL links in them. This presents a problem
for doing this kind of mapping. For example,
suppose both toys.com and cars.com both have a /gifs
directory for storing images, the algorithm needs to
be able to differentiate between the 2 requests.

Here is a scenario that illustrates the problem:

1) On the first request, the browser requests:
http://acme.com/toys

2) The URL is mapped to http://toys.com by the
proxy, and the content is retrieved from toys.com by
the proxy and sent to the browser

3) The returned HTML file from toys.com contains
the following data:

<HTML>
<BODY>
<IMG SRC="/gifs/logo.gif">
Toys Web Site
...

</BODY>
</HTML>

4) The browser now attempts to load the logo image
file by requesting: http://acme.com/gifs/logo.gif

5) The proxy does not know where to get the image
file since it could have been referenced from any of

2

Page 3 of 6

the 3 Web sites (i.e. toys.com, cars.com, plants.com

)

Solution

This algorithm solves the problem by remembering
information about the previous request so that it
can retrieve subsequent requests from the
respective We...