A hybrid algorithm based-on Gossip architecture by using SVM for reliability in cloud computing

Nowadays, Cloud computing is very important due to its widespread use. Significant cloud computing expansively and flexibility, along with other benefits cause to be a background of creating reliability challenges and this kind of challenge is felt more because of the numerous users. To solve this challenge in recent years, many researchers have provided a variety of fault tolerance algorithms to deal with and fix faults in cloud computing, but despite many efforts, unfortunately, it has not been able to solve all the problems. The purpose of this paper is to present an efficient and new hybrid combination using special SVM; at first with using SVM virtual machines are separated and categorized based on their behavior patterns, and then using the Gossip Protocol, the operation of the case for each group and cluster is based on behavioral patterns and in order to achieve QOS it is better to predict. Conclusion of performed simulation in Cloudsim shows that recommended method increases processing speed around 0.65, and makespan is reduced at 7.22s by new method than similarity methods.