Research on efficient allocation of high-level remote sensing image resources and services under cloud environment

Research on efficient allocation of high-level remote sensing image resources and services under cloud environment

Zhejiang University 曾志偉

The specific research contents include the following aspects: 1 based on the analysis of Grid computing and cloud computing in the advantages of high performance computing, a grid computing and cloud computing Fusion Strategy is proposed to solve the high efficient remote sensing processing method of large data. Then, according to the WebService specification, the integrated description mechanism of resources and services is studied, especially for the descriptive method of stateful resources, which facilitates the organization and management of resources and services and achieves efficient access. 2 based on the advantages of peer-to-peer in distribution, reliability and fault-tolerance, this paper studies a peer-to-peer Network dynamic resource and service discovery algorithm under the cloud environment, and realizes the choice of high resolution image data processing dynamic computable resource efficiently; From the perspective of enhancing the efficiency of task execution, After the logical expression of the discovered computational resources, a dynamic resource mapping algorithm based on independent task processing is proposed to realize the formation of the virtual space of computational resources. 3 based on the analysis and discussion of the demand allocation and grid computing economy model of cloud resources, this paper proposes a new multi-objective utility optimal resource and service provisioning model GL-RSPM for the user's task request. 4 a disaster test system was constructed on GL-RSPM. The system architecture is designed and the hardware and software are deployed. The test system realizes the registration service of resource and service, Meta Service Management Service, workflow and QoS setting service, high resolution remote sensing image data processing and so on. According to the related algorithm, the experiment shows that the algorithm is correct and the performance is reliable. The research shows that, from the point of view of resource and service management and deployment, this paper puts forward an integrated description method of environment resources and services based on cloud Grid computing, Peer-to-peer Resource and Service discovery algorithm, and high efficient allocation model for massive high score remote sensing images, which can make the users and resource providers achieve the best efficiency. Provides a reference for large data-intensive computing to achieve efficient provisioning of resources and services.

Keywords-Wenger (Grid computing + cloud computing) SOA High-resolution remote sensing image Peer-to-peer high Performance computing

[Download Address]:http://bbs.chinacloud.cn/showtopic-14354.aspx