Abstract:
JDCP project aims at developing a platform for creating high performance computing clusters. It aims at exploiting the existing hardware resources such as PCs and normal networking and combines them to create a single larger virtual processing machine. The initial distributed computing system was available before the start of the project. This basic platform enabled simplest distributed computing. During the course of the project new features were added to this existing system. The features added include cluster monitoring, adaptive real-time scheduling, information services, fault tolerance, batch job support and certain optimizations to reduce network data transfers. Detailed discussion on these features constitutes a major part of this report.
Firstly the report gives a brief account of the initial system. Afterwards a comparison study is given that compares the JDCP system to existing clustering systems available and discusses how JDCP differs in many aspects of a distributed computing system. Once the significance of JDCP is explained, the report discusses in detail the features added into the system during the course of the project. This discussion not only includes the work done but also includes the need and benefits of these features. In the end of the report testing results and a brief discussion on them is given and on the basis of these results a concluding discussion is done on the whole project.
