Raising the level of abstraction of application-level checkpointing
Ritu Arora · 2008
Checkpointing is one of the key ingredients required for writing fault-tolerant applications for dynamic and distributed computing environments. Manual reengineering of large legacy applications to insert checkpointing mechanisms is a daunting task, mainly due to the time and cost overheads. This poster describes an approach for nonintrusive reengineering of existing applications to insert a Checkpointing and Restart (CaR) mechanism. The user describes the CaR specifications at a high-level and the required code is semi-automatically generated and inserted in the application.