RIRM - Fault tolerant execution incorporating ROC for multitier transactional applications
C. Selvakumar, M. R. Sumalatha · 2013
Effective utilization of resources in cloud computing is a challenging issue in cloud computing. Failure can happen, when there is concurrent execution of cooperating processes. Currently, redundancy mechanisms are implemented to deal with this problem. This reduces the chance of achieving appropriate utilization of the computing system. The proposed RIRM - (Fault tolerant execution incorporating ROC for multitier transactional applications) system focuses on providing better system utilization by supporting parallel processing which is failure proof for dependency based failures in Delta enabled cloud services. Delta enabled cloud services store incremental changes / updates in the cloud environment. In RIRM secure data storage and recovery mechanisms are provided for improving the reliability of services. Dependent processes are identified by the log files and process connectivity graph is constructed. Any failed process in the process connectivity graph is either recovered or redundant processes will be taken in order to enable uninterrupted execution. Process diagnostics is provided by recovery oriented computing that focuses on recovering the failed process which has one or more nodes that are dependent on the considered process. By this approach, failure occurrence has been efficiently identified and process utilization has been improved.