Dynamic and scalable virtual machine placement algorithm for mitigating side channel attacks in cloud computing

Abstract

Abstract Sharing of resources over the cloud computing network tends to the leakage of authorized information of the victim machine by allowing a malicious machine to mine the delicate information from a co-resistant machine known as side-channel attacks. NOMAD is one of the algorithms introduced to defend against the side-channel attacks by minimizing the co-residence between the virtual machine using the greedy baseline algorithm. This baseline greedy algorithm is enumerated with a large number of the set of moves that considered entire possible N-ways or pair-wise swaps between the virtual machines. Thus, re-computing the benefits for every move is computationally expensive. To address this issue, we proposed a dynamic and scalable virtual machine placement algorithm, where the computation of the set of moves is limited by using the Previously-Selected-Server-First (PSSF) Policy. From the obtained selected set of moves, the greedy baseline design is implemented. In PSSF, the highest priority is given to servers that are previously hosted or already hosted virtual machine from a similar user, whenever a new virtual machine placement request is being processed. The performance comparison of this approach is carried out using 20 different Virtual Machines in 4 different servers and showed that it has better efficiency evaluations such as hit rate, loss rate, and resource loss when compared to other VM placement policies.