A practical scheduling algorithm to achieve 100% throughput in input-queued switches

Abstract

Input queueing is becoming increasingly used for highbandwidth switches and routers. In previous work, it was proved that it is possible to achieve 100% throughput for input-queued switches using a combination of virtual output queueing and a scheduling algorithm called LQF. However, this is only a theoretical result: LQF is too complex to implement in hardware. In this paper we introduce a new algorithm called Longest Port First (LPF), which is designed to overcome the complexity problems of LQF, and can be implemented in hardware at high speed. By giving preferential service based on queue lengths, we prove that LPF can achieve 100% throughput.

Parallel desynchronized block matching: A feasible scheduling algorithm for the input-buffered wavelength-routed switch

“…The global optimum to the matching problem defined by (1) can be achieved by implementing the modified Edmonds-Karp Maximum Size Matching (MSM) algorithm proposed in [13]. This algorithm obtains the matching of the maximum size, giving preference to the arcs destined to lower delay nodes, when performing path augmentation [12].…”

Scheduling reserved traffic in input-queued switches: new delay bounds via probabilistic techniques

“…We remark that we cannot directly compare the expressions (5) and (8) for the Random Permutation and Poisson Competition schedulers with the bounds (4), (6) and (7) considered in this section. This is because the expression (5) …”

“…Analyses of MWM-type schedulers can be found in, for example, [7], [8], [9], [10], [11]. Some frame-based schedulers were presented in [19], [20].…”

“…These schedulers are sometimes known as maximum weight matching (MWM) type schedulers (e.g. [7], [8], [9], [10]). Note that an MWM scheduler must operate in real-time since it needs access to queue information and hence it must have extremely low complexity.…”

CIXB-1: combined input-one-cell-crosspoint buffered switch

“…Maximum matching algorithms are efficient but with such a high complexity [2] that implementation is infeasible for high-speed systems. Schemes based on a maximum size or weight matching, like Longest Port Queuing (LPQ), Oldest Cell First (OCF), and Longest Port First (LPF) [3], have been proposed [2].…”