Introduction

The World Wide Web (WWW) is experiencing exponential growth (in number of users, Web servers, multimedia content, real-time audio/video transmissions), causing a rapid increase in Web traffic over the internet infrastructure. The number of users - both individuals and businesses - is growing at a rapid pace. This growth in HTTP traffic poses scalability problems. Network pipes get bigger with the new technologies, making users access more data during the time they are connected, the content of the Web documents includes more and more multi-media capabilities. In corporations more employees are telecommuting. This growth is expected to continue at a rate that addition of networking infrastructure cannot keep pace with.

This in all implies problems as

• increased network bandwidth usage
• increased latency
• increased load on document servers

In these conditions, unless solutions are found for reducing the network bandwidth usage, the latency of document access and the demand on document servers, the Web access will become slower and slower, while the bandwidth costs will increase disproportionately.

Reducing the latency on the Internet Several techniques have been proposed for reducing the latency on the Internet. However, they are all based on the same idea: latency, bandwidth demand and server load are all reduced by migrating copies of popular documents closer to the user.

There are two models of migration according to which side initiates the migration:

1.

replication, initiated and controlled by the server and

2.

caching, which is automatically done in response to user requests.