Non-homogeneous hidden Markov model approach for load balancing in web server farms (NH2M2-WSF)

The web server farm usually consists of many servers. These servers are in the different states at the different period of time. The process of load balancing plays a vital role in the web server farm. The objective of this paper to balance the load equally among the servers present in the web server farm. To relocate the load among the servers, each server´s state should be known. The hidden markov model is used to know the states of the server theoretically and these states are hidden from the clients. The markov chain is used for the transition of the server´s state. It defines the each server´s state and the next corresponding state. Thus a theoretical model is derived for various states of the server and load balancing process is carried on.