(A) rollback recovery scheme for fault tolerant applications and its implementation

Abstract

In this thesis, a method of rollback recovery in an operating system is presented and an approach to coordination method of recovering concurrent processes for fault tolerant processing is designed and implemented in an operating system. Hardware reliability is improved greatly with the advent of the semiconductor technology but software faults have become increasingly prevalent with the steadily increasing size and complexity of software system. This thesis adopts an approach of rollback with recovery block for the concurrent real-time system. Recovery function in the operating system supports structured incorporation of processings by special recovering actions. The operating system includes several functions to make recovery blocks and to manage concurrent processes. The cost of the rollback recovery scheme is measured by processor time and memory space which are required for saving process states and the results are discussed.