ISSL’s paper accepted to IEEE Transactions on Services Computing
2020.12.02
A research paper entitled “SDRP: Safe, Efficient, and SLO-Aware Workload Consolidation through Secure and Dynamic Resource Partitioning” has been accepted for publication in IEEE Transactions on Services Computing. The paper is co-authored by Myeonggyun Han and Prof. Woongki Baek at Intelligent System Software Lab. (ISSL), CSE, UNIST.
This work proposes SDRP, secure and dynamic resource partitioning for safe, efficient, and SLO-aware workload consolidation. SDRP dynamically partitions the hardware resources between the latency-critical and batch workloads consolidated on the same physical server system and performs resource sanitization to prevent any information leakage across different security domains. Through quantitative evaluation, this work demonstrates that SDRP achieves high resource sanitization quality, introduces low performance overheads, delivers high resource utilization with the service-level objective (SLO) and security guarantees, and defeats the last-level cache (LLC)-based side-channel attack.