Research On Cloud Workflow System And Its Spatial Application

Accompany with the development of information service and rapid expansion of cloud computing, scientific workflow system is facing challenges of growing size of heterogeneous data, complexity of scientific computing and difficulty of task integration. In traditional paradigm, a large number of high-performance computing (HPC) infrastructures were purchased, which would be idle in low peak. Management of mass heterogeneous distribution data, compute-intensive applications support and resource utilize are problems for workflow system to be solved.In this paper, a high-performance, scalable, distributed, parallel workflow system CSWf is proposed. A simple but effective cloud service workflow modeling language CPDL is designed, and a MapReduce compatible cloud workflow engine is developed. We then utilize an efficient virtualized parallel computing model for CSWf. Research realizes urban air pollution workflow with different input data size. Application result demonstrates that CSWf can significantly improve the efficiency of geo-workflow execution.Main contents and contributes of this thesis are as follow:(1) Heterogeneous data of geo-workflow is rapid expansion, and task integration between systems become complex. Based on this current situation, an overview on the research of cloud computing and workflow both home and abroad was given. We then propose the target, content and technical route of this thesis.(2) Basic theory of cloud computing and workflow were expounded, and spatial information service was added to research. Cloud service model, strategy and sercurity was studied, Web services, geo-processing service as well as their construction standard were analyzed. Besides, we also have stress on the research of workflow specifications, modeling method and process definition language, analyse the adaptability of these technologies. Based on these researches, data storage and processing strategy of cloud workflow as well as implementation of geo-workflow were proposed according to the special nature of spatial information and spatial data.(3) According to the specificity of cloud computing and spatial data, we design a scalable, distributed, parallel workflow system CSWf, which utilze the independent design cloud workflow process definition language CPDL and realize a high performance cloud workflow engine with virtualized parallel computing model. CSWf can well settle problems of workflow modeling, scheduling, coordination and execution between distributed workflow systems. Research realizes urban air pollution workflow with air pollution data released by the U.S. Environmental Protection Agency. Compare with traditional workflow software, CSWf has performance advantages in dealing with large amount of data.