Specialized N-Modular Redundant Processors in Large-Scale Distributed Systems

Computers are being used to achieve increasingly sophisticated control for large and complex systems. Many of these systems require a large shared state space or database. Thus, handling real-time concurrent accesses to a shared database becomes an essential feature for modern fault-tolerant systems. Many fault-tolerant systems have been implemented for uniformly tolerating various types of failures, such as MAFT, FTP, FTPP, and Delta-4. However, most of these either lack the notion of a shared state space or do not eeciently support parallel tasks that concurrently access a shared state space. We use a processor-specialization approach to increase the eeectiveness of replication and, consequently, achieve cost-eeective fault tolerance in such systems. The SNMR protocol is developed based on these concepts. Compared to many existing Byzantine-resilient systems, the SNMR approach incurs less overhead and can be easily pa-rameterized to t various fault models.