Research On Key Technologies For Cloud-Oriented Scientific Workflow

Scientific computing refers to the application of computers to solve scientific problems encountered in the scientific research and engineering process. As a theory, the experiment equally important third scientific research means scientific computing in many fields of scientific research, such as the field of high-energy physics, weapons development, bioinformatics, atmospheric science plays a very important and irreplaceable role. Compared to modern scientific research in traditional scientific research, large-scale cooperation, horizontal and vertical linkages between the multidisciplinary nature of disciplines more closely; methods and means of modern scientific research is increasingly dependent on emerging information technology, network technology; modern scientific research process usually requires the combination of different fields, the application of multiple organizations found, processing, storage and visualization of massive data operation. Traditional single field of science-oriented, scientific problem solving environment of a single organization or project can not meet the needs of modern large-scale multi-disciplinary scientific computing, the need for a collaborative scientific computing environment for multidisciplinary provide users with a data sharing and collaborative work virtual space, in order to achieve a collaborative interdisciplinary process of complex scientific computing.Workflow as a process automation and collaboration technology in the field of modern enterprise business process management, e-government and e-commerce has been a wide range of research and application. Scientific Workflow is automatic or semi-automatic workflow technology in the scientific research, scientific workflow is solving a scientific problem solving process computer implementation, application and development of workflow technology in scientific computing. Traditional scientific workflow for specific areas and research projects, calculate the required data and resources are stored and distributed in the organization-wide, can not support the use of distributed, heterogeneous platform computing tools and resources. Grid workflow technology is developed in order to effectively solve the problem of the automation of the operation of the flow characteristics in grid and scientific computing, grid services operating flow, these grid services collaboration to complete a specific goal. However, the grid workflow systems usually need to perform complex distributed computer systems, such as high-performance computers, distributed cluster system and grid system, construct such systems often require to pay unusually expensive price. To apply for access to these systems also require complex and time consuming process.Cloud computing is distributed computing, parallel computing and grid computing’s further development, it is the provision of Infrastructure as a Services (IaaS), Platform as a Services (PaaS) and Software as a Services (SaaS) Internet-based computing to a variety of Internet applications. Cloud computing provides shared infrastructure, it will be distributed through virtualization technology, virtual computing resources and storage resources in different geographic locations into a pool of resources, application resources, users need to use the resources to use after the release of resources, and to provide resources reusability. In this way, the cloud computing center can provide high-performance, low-cost and simple to use computing resources and vast amounts of storage resources. With cloud computing in-depth study and promote the application of the technology, its highly efficient, flexible, on-demand service, pay and quality characteristics of a new way of thinking to solve problems in scientific workflow running. Cloud-based computing environment, cross-organizational automation of of workplace synergy of interdisciplinary data processing and scientific research process has become one of the problems to be solved in the scientific research.Scientific workflow technology and applications in cloud computing environment is becoming an urgent research spot of the scientific workflow field. The dynamic and open nature of cloud computing runtime environment, the distributed autonomy of cloud services and the loosely couple of scientific workflow’s logic, make cloud-oriented scientific workflow systems can give full play the comprehensive advantages of both, however it proposes a new problem for scientific workflow research in cloud computing: how to creatively establish a fault-tolerant cloud-oriented scientific workflow model, and how to effectively improve the quality of service(QoS) of cloud-oriented scientific workflow. Based on the existing works of the fault-tolerant mechanisms and scientific workflow scheduling mechanism in cloud computing technology and scientific workflow development methodology, the paper designed a new fault-tolerant cloud-oriented scientific workflow system architecture description language CSWFADL-EH, and proposed a cloud-oriented scientific workflow QoS computational model and management system framework CSWFQoSM, and come up with the corresponding algorithm established; then The CSWF-F framework is designed to support fault tolerance and QoS management of cloud-oriented scientific workflow, at last, based on the Geological disaster warning scientific workflow, discusses the prototype system of CSWF-F, including the system supporting environment, as well as key implementation techniques.