ABSTRACT

This paper will illustrate the associated issues with multi-tier enterprise applications in the cloud when the tiers need to distribute geographically as in multicloud and hybrid cloud. The end-to-end response time for such application will largely depend on network delays as the control moves from one tier to the other tier. An effort made by proposing a simplistic deployment pattern to mitigate the delay in response time by minimizing the network communication between geographically distributed nodes of inter-tier clusters. A deployment pattern proposed in paper attempts to combine the functionality of specialized nodes of different tiers into a more generalized, multi-purpose node cluster to the possible extent. A formalization of the gain in response time based on the experiment incorporating three deployment models of the cloud for a web application. A detailed analysis presented at communication time of the different deployment model and results used to understand and draw conclusions to determine the benefit of proposed deployment pattern.