An optimistic technique to detect Cache based Side Channel attacks in Cloud

Abstract

Data security is the most critical field in Cloud Computing. The Critical data can leak through unpredictable side channels, posing very risky threats to information security. The Cache-based Side Channel Attacks (CSCAs) is one of the most challenging attacks. Different secure cache architectures have been proposed to defend against these attacks. But these solutions are not reliable to detect and prevent those attacks. Detection of CSCAs is a very important research-related problem. In this paper we demonstrated the detection of CSCAs. We measured the vCPU cycle, virtual memory utilization and cache miss rate to detect the attackers. We illustrated Prime + Probe attack, Flush + Flush attack, and Flush + Reload attack on the AES cryptosystem. Our detection technique is compared with the existing detection solution to demonstrate the accuracy. Our proposed work is achieved 92.5% accuracy to detect the CSC attacks. With the help of our proposed system the Cloud Service Providers(CSPs) can identify the attackers of VMs and host machines to safe guard victim users from the CSCA.