A Quantitative Comparison of Scheduling Algorithms for Input-Queued Switches

This paper quantitatively evaluates several alternative approaches to the scheduling of cells in a highbandwidth input-queued ATM switch. In particular, we compare the performance of three algorithms described previously: FIFO queueing, parallel iterative matching (PIM), maximum matching and two new algorithms: iterative round-robin matching with slip (SLIP) and least-recently used (LRU). For the synthetic workloads we consider, including uniform and bursty tra c, SLIP performs almost identically to the other algorithms. Cases for which PIM and SLIP perform poorly are presented, indicating that care should be taken when using these algorithms. But, we show that the implementation complexity of SLIP is an order of magnitude less than for PIM, making it feasible to implement a 32x32 switch scheduler for SLIP on a single chip.