Recovery blocks in real-time distributed systems

This paper discusses the concept of recovery blocks as a dynamic redundancy approach to software fault tolerance. The discussion focuses on the distributed recovery block (DRB) scheme which can be thought of as a means of integrating hardware and software fault tolerance in a single structure. The DRB approach, which combines distributed processing and recovery block concepts, is capable of effecting forward recovery while handling both hardware and software faults in a uniform manner. The DRB was developed for applications such as command and control in which data was collected by interface processors and distributed over a network, and in which data from one pair of processors was output to another pair of processors. The extended distributed recovery blocks (EDRB) is then discussed as a modified scheme of the original DRB for real-time process control applications. The implementation of the EDRB is also presented to acquaint the reader with the implementation issue that must be faced in the development of a fault-tolerant software architecture for a distributed system.