Fast and Noniterative Scheduling for Input-Queued Switches with Unbuffered Crossbars

Most high-end switches use an input-queued or a combined inputand output-queued architecture. The switch fabrics of these architectures commonly use an iterative scheduling system such as iSLIP. Iterative schedulers are not very scalable and can be slow. We propose and study a new scheduling scheme that is fast and scalable. This scheduling scheme finds a maximum matching of a modified I/O mapping graph in a single iteration. Analytically and experimentally, we show that it provides full throughput and incurs very low delay; it is fair and of low complexity; and it greatly outperforms traditional iterative schedulers. It also renders arbitration egress memory unnecessary. We discuss the issues related to implementing our scheduling scheme on a conventional crossbar switching fabric. Finally, we present a switch architecture that is well suited for implementing cell switching based on our noniterative single-iteration scheduling scheme. Our scheduler, its underlying queuing structure and arbiter circuit, and its supporting interconnection circuit form a complete switch fabric architecture that is simple and effective.