A Classification Approach for Proactive Fault Tolerance in Cloud Data Centers

摘要

Reliability and availability are the two greatest challenges faced by cloud data centers as it affects the service level agreement between the provider and consumer. Most of the existing approaches use checkpoint and reactive mechanisms which are not considered to be efficient due to performance issue and challenges. So an improved fault tolerance technique is more under research to solve the issue of fault tolerance. This paper proposes a Fuzzy min max neural network Classification approach that can predict failure in advance and solve the problem in case of overlapping. Bu initially training the data set and the threshold value is fed into the testing data set using the proposed Fuzzy min max neural approach. Results show that there is an improvement in the performance when compared over the previous approaches.