Publication

Abstract : Cloud computing offers on-demand access to a large pool of shared resources at lower cost. The advantage of cloud resources is that it can be easily provisioned, configurable, and managed with minimal management efforts by the users. Proper load balancing is an important task in maintaining fault tolerance and Quality of Service (QoS). In the cloud, a load balancer accepts incoming user requests, application specific traffic and distributes this workload across multiple backend processes using various methods. In a single load balancer system; if the load balancer is down none of the user tasks can’t be processed, even when the servers are ready to process the tasks. In order to overcome this single point of failure, this paper proposes a model that will avoid the single point of failure by using multiple load balancers. In this method, service of one load balancer can be borrowed or shared among other load balancers when any correction is needed in the estimation of the load. This will improves fault tolerance of the cloud eco system and assist in cluster capacity management.