CryptVMI: Encrypted Virtual Machine Introspection in the Cloud

Abstract

Virtualization techniques are the key in both public and private cloud computing environments. In such environments, multiple virtual instances are running on the same physical machine. The logical isolation between systems makes security assurance weaker than physically isolated systems. Thus, Virtual Machine Introspection techniques become essential to prevent the virtual system from being vulnerable to attacks. However, this technique breaks down the borders of the segregation between multiple tenants, which should be avoided in a public cloud computing environment. In this paper, we focus on building an encrypted Virtual Machine Introspection system, CryptVMI, to address the above concern, especially in a public cloud system. Our approach maintains a query handler on the management node to handle encrypted queries from user clients. We pass the query to the corresponding compute node that holds the virtual instance queried. The introspection application deployed on the compute node processes the query and acquires the encrypted results from the virtual instance for the user. This work shows our design and preliminary implementation of this system.