A Case for Scaling Applications to Many-core with OS Clustering

摘要

This paper proposes an approach to scaling UNIX-like operating systems for many cores in a backward-compatible way, which still enjoys common wisdom in new operating system designs. The proposed system, called Cerberus, mitigates contention on many shared data structures within OS kernels by clustering multiple commodity operating systems atop a VMM, and providing applications with the traditional shared memory interface. Cerberus extends a traditional VMM with efficient support for resource sharing and communication among the clustered operating systems. It also routes system calls of an application among operating systems, to provide applications with the illusion of running on a single operating system. We have implemented a prototype system based on Xen/Linux, which runs on an Intel machine with 16 cores and an AMD machine with 48 cores. Experiments with an unmodified MapReduce application, dbench, Apache Web Server and Memcached show that, given the nontrivial performance overhead incurred by the virtual ization layer, Cerberus achieves up to 1.74X and 4.95X performance speedup compared to native Linux. It also scales better than a single Linux configuration. Profiling results further show that Cerberus wins due to mitigated contention and more efficient use of resources.