On Time Rendering with Failover of Tasks in Multicloud Clustering

Abstract

Cloud computing is an abstraction for the complex infrastructure, computation and software management over a network, a way to rise up the capacity or potential on the fly which entrusts services to a heterogeneous community. Multicloud offers a management platform for delivering dynamically-scalable services, computing power, bandwidth and storage to the public as utility computing without any bound for time and location. The proposed system focuses on a multicloud cluster with failover and scaling ability, in which the cluster nodes deployed on multicloud are configured to provide seamless connectivity between clouds. In this work, feasibility of the multicloud deployment spanning private and public cloud is analyzed. Failover is an automatic switching of task from one node to other in case of failure. The status of node is recognized by the sharing of constant heartbeat messages and it leverages checkpoint functionality to overcome failures. The clustered nodes deployed on different clouds can maintain built in automatic failover between different cloud providers, can improve the efficiency through scaling and helps to increase the throughput using zero deployment modes. This work assures the reliability of adopted multicloud system with universal control.