Research On Monitor System Of Volunteer Computing Platform And Its Application In Power Flow Computation

Volunteer computing has been proved a promising idea to aggregate idle CPU cycles widely available on Internet for large scale and distributed applications. Although current volunteer computing projects in the world have been proved successful, there are still some problems left to solve. These problems include: complicated API (Application Programming Interface) to make application developing non-trivial, lacking a friendly user interfaces, specific application-oriented, single point of failure, inscalability and etc. Considering these problems, a new cooperative server group based volunteer computing platform called P2HP, which overwhelms the disadvantages of current volunteer computing projects is presented in this paper. P2HP is open, ease of use, fault tolerant, scalable and platform independent. A convenient API (Application Programming Interface) set is also provided to support the development of parallel applications running on P2HP.Monitoring system is an important module of P2HP. Queue based approach that is easy to implementation, is adopted to manage membership such as nodes joining or leaving. By utilizing object serialization, P2HP is platform independent. P2HP can run under different kinds of environments which are composed of different kinds of hardware and software. A highly efficient message communication mechanism and"heartbeat"technique are designed it meet the communication demands among the different kinds of Monitor nodes. By using this mechanism, the monitor system can monitor the nodes in its own group and the subordinate Dispatcher group efficiently. The information of the application running on the P2HP can also be monitored by Monitor system.Power System is a wide area distributed System. The power flow computation, which is computing-intensive, coarse parallel granularity, and low bandwidth requirements, can be supported by P2HP. According to the characteristics of power flow computation and P2HP, An adaptive task partition for the power flow computation on P2HP is described according to their characteristics. Contingency in power system is utilized to decide partition criterion. By using the API (Application Programming Interface) provided by P2HP platform, constructing the high performance computing programming environment, setting up the user work space, writing a Main program and Sub program, Power flow computation application was developed on P2HP. Then several implementation technologies are described in detail. The Monitor system can monitor the running power flow task.The function testing of Monitor system shows that Monitor system of P2HP can provide completed functions which include node management, task management and application monitoring. The performance testing of power flow application use standard benchmark, IEEE14 node power system, the results demonstrated that P2HP can meet the demands of real time and secure power flow computation in large scale electric power systems. A good speedup is also obtained in this test.