Fault Recovery in Mesh type Multiprocessor Networks

In a mesh type multiprocessor network, failure of a single node or link may change the topology of the network and at worst partition it into several disjoint parts. This may prevent any recovery technique to be implemented. Using redundancy techniques it is possible to keep a network connected even after a fault occurrence. If topology of a network is kept intact after a fault occurrence, not only the connectivity of nodes of the network, but also the performance

کلیدواژه ها:

Fault- tolerance ، Recovery ، Multiprocessors ، Fault-tolerant graphs ، Reconfiguration