Multistage network with an additional stage for fault tolerance

The extra stage cube (ESC) network, a fault tolerant structure, is proposed for use in large-scale parallel and distributed supercomputer systems. This network is derived from the generalised cube network by the addition of one stage of interchange boxes and a bypass capability for two stages. It is shown that the ESC provides fault tolerance for any single failure. Further, the network can be controlled even when it has a single failure, using a simple modification of a routing tag scheme proposed for the generalised cube. Both one-to-one and broadcast connections under routing tag control are performable by the faulted ESC. The effects of the extra stage on the partitioning and permuting abilities of the network are described. 19 references.