Skip to Main Content
This paper discusses the support for two different implementations of resilient Logical Machines. LM checkpoint/rollback and replicated LM execution, corresponding to backward and forward error recovery respectively. By LM approaches, one can effectively achieve fault-tolerance for node crash but eliminate some critical problems such as domino effect, live lock, and local dependencies
Singapore ICCS/ISITA '92. 'Communications on the Move'
Date of Conference: 16-20 Nov 1992