Design And Implementation Of Grid System For Kinetic Disaster Simulation Task In Engineering Calculation

With the constant improvement and development of science and technology, kinetic disaster simulation tasks in the field of engineering calculation become more complex; the construction of task models become more complicated, the code size become larger, and so on. These make the rising demand in equipment performance for task executing, and a single machine or cluster is no longer meets the needs of running tasks. Defined as a distributed system, grid can coordinate distributed resources and provide the best quality of service for Grid users through standard, open and common protocols and interfaces.Based on grid and web services technology, a scientifically-oriented Grid system was proposed with major engineering calculations as its background. In addition, the system functionality was designed and its operating process was implemented too, which effectively reduced the running time and improved the utilization efficiency of software and hardware resources.Grid service is the core of grid technology. SOA and web services have been applied to grid architecture in order to improve collaboration mechanism in heterogeneous grid environment and bring benefits to the scalability and interoperability of grid systems. Moreover, it enables grid services in heterogeneous distributed environment to have a better collaboration and dynamic combination capabilities. Obviously, it brings new vitality to the development of grid technology. This system focused on the application and operating principle of services for kinetic disaster tasks of major engineering projects, described all stages from service request to service invocation, and gave detailed service implementation and deployment details.Generally speaking, there are two ways for service invocation. A common way is through browser-based portal, which provides a unified approach for service access and invocation. In this way, grid users can easily achieve the submission, operation and results reviewing of scientific computing tasks in the portal. The other way is integrating services into users'own system to meet the needs of the actual task executing. It will bring great convenience to users.Finally, several sets of disaster simulation test cases were executed to verify the functionality of service encapsulation and invocation, and the computing performance was briefly analyzed.