A Distributed Process Management Model for Better Scalability on Multicore Platform

摘要

With the widespread utilization of mul-ticore processors, operating systems face new challenges on scalability. The lock mechanism used by current mono-lithic OSs makes some critical modules encounter perfor-mance bottlenecks caused by lock contention. The con-tention in Linux process lifecycle management results in process Creation, execution and termination (CET) scal-ing poorly with core count increasing. Differing from the centralized fashion of Linux, this paper presents a Dis-tributed process management model (DPMM) on micro-kernel for better scalability. The kernel functions are dis-tributed into m essage passing servers running on differ-ent cores. DPMM splits the shared data including Process control blocks (PCBs), page pools and buffer caches into these cooperating servers. The actions including PCB cre-ation/deletion, page allocation/freeing, executable loading and scheduling can be parallelized without locking. Our micro-benchmarks show that DPMM achieves nearly lin-ear performance growth on process CET and outperforms Linux and Minix3 on a 32-Core m achine.