Log Based Web Server Recovery System

Abstract

With the explosive growth of consumer and business demand for global Internet, corporate intranets and business-chain extranets use of solutions with client server architecture have grown up. A traditional standalone (non-clustered) server does not provide any failover capabilities. The Log Based Web Server Recovery system discussed in this paper is basically an approach of providing failover capabilities up to a certain extent in a non clustered client server environment. In a client server environment if a server goes down, all its session data is lost since servers maintain sessions in memory. A more reliable solution to overcome this problem would be to replicate session data across several nodes in a clustered manner, so that whenever a server goes down, another server in the cluster can take over rest of the work since session is replicated among the cluster of servers. Even though this is a good solution for high end business users, this is a very expensive solution when it comes to small scale business users since, need of more resources. The Log based Web server recovery system introduced in this paper will persists session data for each requests done by clients in a data base table. Session is replicated and is stored in a table in the database along with its session Id. Even though server fails unexpectedly, user’s session data along with their session ids will persist in the data base accordingly. When the server is back online, when we send the same request, it will restore the swapped sessions from the database back to the client. In this case server will not need to get the clients request again and start work from the beginning. Clients users will not have to re-enter all the data which they have already entered before the server crash. This is a less expensive solution since no additional servers will be used.