ZOMA: a preemptive deadlock recovery mechanism for fully adaptive routing in wormhole networks

Deadlock is the most formidable obstacle that adaptive routing algorithms in wormhole switched interconnection networks must address and overcome. The class of True Fully Adaptive Routing algorithms represents an efficient routing technique with favorable performance and operational characteristics. These algorithms route packets without imposing any routing restrictions. This paper proposes a true fully adaptive routing algorithm with a new deadlock-recovery mechanism. Our proposed deadlock-recovery mechanism, ZOMA, is efficient and takes advantage of the concept of wormhole switching in terms of low hardware resource requirements. The performance of the new mechanism can match that of other more expensive deadlock-recovery mechanisms, while requiring lower hardware resources that are not on the critical path of the switching process. The proposed mechanism creates a new category of deadlock recovery techniques that we refer to as preemptive as opposed to the existing progressive and regressive categories. Performance evaluation of the proposed mechanism against other routing algorithms is accomplished via simulation.