Credit-Based Scheduling Algorithms for Input Queued Switch

概要

論文の詳細を見る
The input queued (IQ) switching architecture is becoming an attractive alternative for high-speed switches owing to its scalability. In this paper, three new algorithms, referred to as the maximum credit first (MCF), enhanced MOF (EMCF), and iterative MCF (IMOF) algorithms, are introduced. Simulations show that both MCF and IMCF have similar performance as the Birkhoff-von Neumann decomposition (BVND) algorithm, which can provide cell delay bound and 100% throughput, with lower off-line computational and on-line memory complexity. Simulations also show the fairness of MCF is much better than that of BVND. Theoretic analysis shows that the EMOF algorithm has a better performance than MCF in terms of throughput and cell delay with the same complexity level as MGF. Simulation results indicate the EMCF algorithm has much lower average cell delay and delay variance as compared to the BVND algorithm.
社団法人電子情報通信学会の論文
2002-09-01