The response time to a web request depends upon the distance between the web client and the web server, the load of the web server, and the network traffic. Replicated web serves can reduce the response time of web requests requesting not only static ...
The response time to a web request depends upon the distance between the web client and the web server, the load of the web server, and the network traffic. Replicated web serves can reduce the response time of web requests requesting not only static documents but also dynamic documents such as CGI scripts. They also increase the availability of web servers. the number and locations of the replicated web servers should be determined considering both how much web requests are generated and where they are coming from. In this paper we proposed a method of dynamically replicating web servers. When the load of a certain web server becomes too high, a new web server is created on the Rent-a-Server which is close to the domain from which heavy requests are coming from. When the load of a web server on a certain Rent-a Server becomes very low, the web server is removed from that Rent-a Server. We also introduced a method which selects the web sever that will service a user's request considering both the proximity of the web server from the client and the web servers' load.