Performance Isolation for Real-Time Applications on Multicore Platforms using PALLOC and MemGuard

Performance isolation among multiple programs on a multicore platform is difficult to achieve due to contention in the shared architectural hardware resources such as shared caches and DRAM. In this paper, we present a case study—using WebRTC, an open source real-time video conferencing software developed by Google—that compares state-of-art techniques to achieve performance isolation. We first investigate the performance variability of WebRTC in the presence of memory intensive co-running applications. We then compare CPUSET (baseline), PALLOC (DRAM bank partitioning technique), and MemGuard (memory bandwidth reservation technique) in terms of performance isolation of the WebRTC application as well as the overall throughput of the entire system. In this study, we found that PALLOC achieves moderate real-time performance improvements without significantly sacrificing the overall throughput. On the other hand, we found that MemGuard can be configured to achieve near perfect performance isolation for real-time applications, albeit at the cost of significantly reduced throughput for co-runners.