Design and analysis of a hardware-assisted checkpointing and recovery scheme for distributed applications

A checkpointing and recovery scheme which exploits the low latency and high coverage characteristics of a hardware error detection scheme is presented. Message dependency which is the main source of multi-step rollback in distributed systems is minimized by using a new message validation technique derived from hardware-assisted error detection. The main contribution of this paper is the development of an analytical model to establish the completeness and correctness of the new scheme. A novel concept of global state matrix is defined to keep track of the global state in a distributed system and assist in recovery. An illustration is given to show the distinction between conventional and the new recovery schemes