Introduction to Flexible Rollback Recovery In Dynamic Heterogeneous Grid Computing Java Project:
Huge applications execution on Grid or cluster architectures consisting of process nodes produces issues with reliability.
The supply of the issues is node failure and therefore they would like for dynamic arrangement over in depth run time.
Two mechanism is described here one is Theft-Induced Check pointing and another one is Systematic Event Logging.
This paper is used to gain the details about the two mechanisms. It is actually transparent protocol. It can recover multi threaded applications.
It is also used to execute of a data flow graph. It is such kind of protocol which is capable to rectify the problem with faults such as crash fault, and node volatility. Dynamic heterogeneous system is also recovered by these two protocols.
It is especially suitable for any kind of application. There is an existing system for this protocol or application and another one is proposed System and both are described below.
- Architecture of a grid or a cluster is consists many different kind of nodes. This type of architecture can create that many type of problem.
- Nodes failure is the main problem in this type of architecture.
- To overcome this problem MTTF i.e. Mean Time to Failure concept arrives. By this approaches execution time of the application should be tolerance free in grid.
- As we know this paper is based on two protocols, these two protocols can execute the Data flow graph. It can recover the heterogeneous system and multi threaded application.
- It can recover different kind of processors but it is mainly built for application. Though it is low cost but it allows the controlling or bounding.
- The size of the protocol is very small. For this reason maximum work lost by it also very small and enclosed.
Download Flexible Rollback Recovery In Dynamic Heterogeneous Grid Computing Java Project with Source Code.