Reliability Enhancement for Cloud Services - A Survey

Today's IT services are moving to cloud computing environment in order to process client request and provide services effectively. In this case, reliability is a major factor to reduce the load in storage and network resource. In traditional method, it uses fault tolerance, replication of VM and storing checkpoint image in a neighboring server. Replication increases cost for a large system and checkpoint image inaccessibility occurs if any node crashes. It can be rectified by distributed storage of checkpoint across every server. This persistent storage can be made using a heuristic algorithm - an optimization problem. These experiments are verified using a cloud simulator. It does not provide effective results on reliability. Thus, in this proposed system a real cloud environment is chosen - OpenStack to verify experiments and to estimate the performance.