Extending PVM with consistent cut capabilities: Application Aspects and Implementation Strategies
Message passing libraries are now widely used to develop parallel and distributed applications. Despite different services are provided by the available packages, still little support is given for distributed consistent cut, a facility which constitutes the base for higher level services both at system level (e.g. distributed debugging and fault tolerance), and at the application level (e.g. distributed simulation). In this paper we discuss different strategies to integrate a consistent cut protocol in PVM system.
KeywordsControl Message Message Complexity Starter Process High Level Service Virtual Time
Unable to display preview. Download preview PDF.
- 2.Clematis A., Gianuzzi V.: CPVM-Extending PVM for Consistent Checkpointing. Euromicro Workshop on Parallel and Distributed Processing (1996) 67–74Google Scholar
- 3.Clematis A., Deconinck G., Gianuzzi V.: A flexible state-saving library for messagepassing systems. 6th Euromicro Workshop on Parallel and Distributed Processing (1998)Google Scholar
- 5.Deconinck G. et al. Fault Tolerance in Massively Parallel Systems. Transputer Communications 2(4) (1994) 241–257Google Scholar
- 7.V. Gianuzzi. An Optimal Snapshot Protocol for Acknowledged Message Passing, Tech. Rep. DISI-TR-96-6, 1996.Google Scholar
- 9.Jefferson D.R. et al.: Distributed simulation and the Time Warp Operating System. 12th SIGOPS (1987) 77–93Google Scholar
- 11.León J., Fisher A.L., Steenkiste P.: Fail-safe PVM: a portable package for distribuited programming with transparent recovery. CMU-CS-93-124 (1993)Google Scholar
- 12.Wismüller R.: Debugging message passing programs using invisible message tags, in Recent Advances in Parallel Virtual Machine and Message Passing Interface, M. Bubak, J. Dongarra and J. Wasniewski (Eds.), LNCS 1332 (1997), (295–302)Google Scholar