Reseach On Cloud-based Service-oriented Python Parallel Computing

With the high-speed and constant development of technology, large data processing has been becoming more and more common in recent years, such as its application in scientific calculation, engineering project and experimental simulation, etc. For high demands of large data processing and calculating in those fields, requirements of computer performance of the computer which is used to process large data are high. The research has being done not only in the aspect of hardware but also of software. And by doing some technical integration, the advent of the new technology can maximally provide parallel computing service without changing the hardware conditions. Best of all is a parallel computing grid of Computer Architecture, which integrates the distributed computer in the grid and then coordinates to finish the demand of computation, so it meets the computational requirements very well.According to the present problems, as well as technology now exists for parallel computing an in-depth study, conducted a parallel computing platform design. This paper mainly studied by parallel computing, cloud computing, virtualization technology to get the technical basis of the study. Parallel computing nodes will still computing services in the form of cloud services platform available. Through the study of service oriented architecture, understand this architecture provides interfaces that can be accepted cross-platform computing data. And has been designed using the Python client service nodes and compute nodes module, python and have the ability to selfreflexive observation, can call library functions to achieve many other prophecies did not send realize. And these are posted on the parallel computing capabilities on a cloud computing platform virtual machine, and also for testing the service node compute nodes designed to achieve intelligent distributed computing request. Through to the structure system of service system of research, understand that this architecture provides the interface, the interface can accept data calculation of cross platform. And Python was used for the client, the service node and the design of the node module, Python reflexiveand self observation ability, can call library functions in many other languages can not realize the function. Theparallel computing function is released in the cloud virtual machine platforms, and also for design of detection of the computing nodes at the service node, distributedcomputation requests to achieve intelligent.