Research On The Application Of Distributed File System Data Scheduling In G/S Model

Since the 90s of the 20th century, as computer technology, network technology, space technology, communications technology and the unprecedented popularity of the promotion, making the rapid rise of information services, and become the most potential industries in the 21st century. Information industry's role in the national economy more and more important, and this trend continues. According to statistics, is closely related with human life in all of the information, and location information related to about 80%, we can see, spatial information services in the information services industry plays a very important position.Determine the specificity of spatial information services to the special nature of spatial information, spatial information services and these characteristics limit the development of spatial information services, and general information service model and technology architecture can not meet the spatial information services. In recent years, with the national or regional basically completed the construction of spatial databases, on the one hand, massive applications of spatial data for export; the other hand, a large number of civil engineering urgent need for effective spatial data. The key problem is the service model, therefore, the law of spatial information services, spatial information services to explore new models and technology to meet the growing demand for spatial information service application is important, and highly effective meaning.Spatial information services and the further application of the extension, making the information network service is moving in space with high efficiency, low cost, universal simplicity, transparency, and intelligence in real time the main direction of sustainable development, new features of spatial information services and the growing application requirements and available space information service model conflicts between the growing. Grid computing, cloud computing, the concept of a new generation of distributed computing and application made to resolve this contradiction provides a new breakthrough. Information network service in space, when the server-side aggregation service, the network bandwidth and services will generate a huge amount of data network access bottleneck, and when the number of concurrent users increases the number of and access to a certain extent, the efficiency of server aggregation service will decrease.Popular spatial information applications require network services anytime, anywhere, accessibility to pass the required services. The complexity of spatial information services, spatial information distribution of heterogeneity, and the diversity of user requirements, personalized service model makes spatial information networks should have a pervasive and collaboration-oriented features. The so-called universal properties, spatial information is used to store, update, transmission and scheduling of the computing environment can generally adapt to the complex task of hardware and software environment and changing needs; and collaborative features refers to the user environment aggregation of spatial information services to create a collaborative implementation of the distribution system.For the C / S and B / S computing model in the application of spatial information network service deficiencies in the existing mode of spatial information and application network services research, based on the analysis of spatial information services through the information flow and structure characteristics , grid computing and cloud computing will be in the "convergence" concept and the network service model spatial information combined with spatial information is proposed network service model - Geoscience browser / server farms of distributed spatial data (G / S) mode.G / S mode is the essence of Web services for spatial information in the massive, multi-source, multi-resolution, multi-temporal spatial information of the popular application and market requirements, the use of XML-based Geography Markup Language HGML super distribution network of various form a unified description of spatial data and visualization to enhance the geological features of the browser - to learn the browser as a client, through HGML organization, management of distributed spatial data spatial data server group as the server side, through the " request - Aggregation - service "mechanism for client-side aggregation service client dynamic aggregation services, in order to achieve spatial information network services. This thesis originate from the project of the tutor, based on the technology of G/S distributed server clusters, to solve the application requirements of the G/S model of spatial information server clusters as the goal, on the basis of the analysis of the problems that G/S model was meet in provide spatial information services, do the following work:1. An improved algorithm for dynamic file allocation balance. Dynamic file allocation based on the original balance algorithm to add two servers threshold is intended to enable a distributed file system server threshold is set a uniform standard.2. A distributed scheduling algorithm for file system data.This data is based on dynamic scheduling algorithm for file allocation balance, modify the source code Hadoop support of HGML improvethe scheduling of real-time data access, timeliness and accuracy. To meet the G / S mode applications.