A High-Performance Scheduling Algorithm for Buffered Crossbar Switch

Abstract

The limitations in complexity and scalability of current combined input and cross-point queuing switching scheduling policies are first analyzed. To overcome the deficiencies in supporting fair and QOS scheduling, we propose a fair and simple high-performance scheduling algorithm for Combined Input-Cross-point-Queued Switches, which is called Fair Service and Group Smoothed Round Robin (FGSR). The main idea of FGSR is that it first groups flows according to their weight, then decides which group to be scheduled by SRR and finally decides which flow from the selected group to be scheduled. FGSR is based on round robin mechanism, whose complexity is only O(1). FGSR groups and schedules flows according to the weight of flows, thus it has good fairness and can adapt to the need of real-time performance. Theoretical analysis and simulation results show that FGSR exhibits good delay, throughput and anti-burst performance.