Secure Distributed Computing in Cloud Using Trusted Execution Environments
Teodora Radaljac, Danko Miladinović, Žarko Stanisavljević, Pavle V. Vuletic · 2024
This paper investigates the impact of integrating AMD Secure Encrypted Virtualization (SEV) technology into a distributed computing environment. Distributed computing frameworks are designed to handle large-scale data processing and complex computational tasks by distributing workloads across multiple nodes. However, ensuring the security and integrity of data in such environments, particularly in multi-tenant and cloud settings, poses a significant challenge. Confidential computing concepts and in particular AMD SEV address these concerns by providing virtual machine memory encryption and isolation, even protecting against threats from the hypervisor. This study evaluates the potential to set up distributed computing in the confidential computing settings. It will discuss current gaps and the performance overhead of using AMD SEV in distributed systems, focusing on task execution times and scalability. Experiments were conducted using different workloads studying performance impact. Results indicate that while SEV introduces additional data processing, the overall average performance impact is minimal, making it a viable solution for secure distributed processing without significant performance degradation.