An Efficient and Scalable Approach for Implementing Fault Tolerant DSM Architectures

Distributed Shared Memory (DSM) architectures are attractive to execute high performance parallel applications. Made up of a large number of components, these architectures have however a high probability of failure. We propose a protocol to tolerate node failures in two classes of DSM architectures: Cache Only Memory Architectures (COMA) and Distributed Virtual Shared Memory (SVM) systems. The proposed solution is based on backward error recovery and consists of an extension to the existing coherence protocols to manage data used by processors for the computation and recovery data, used for fault tolerance. The implementation of the protocol in a COMA architecture has been evaluated by simulation. The protocol has also been implemented in a \textscsvm system on a network of workstations. Both simulation results and measurements show that our solution is efficient and scalable