The Research Of Software Service Platform Based On Cloud Computing

The software resources are more and more abundant because of the rapid development of information technology, and components are also growing rapid, which is a part of the software resources. In order to use component resources better, different component library are established in 7 cities in China respectively. Unfortunately, the data of the library are hard to share owing to different managements or retrieval techniques. Traditional manage system of software resources are mainly centralized, leading to a trouble that the rich materials while poor product to utilize, that is to say, many resources can't be efficient use. As a result, mass of exploit resources are wasted, and large number of storage resources are occupied simultaneously. Therefore, research on services platforms to unified manage software resources is an urgent problem. A new method, which is called "cloud computing", provides a fitting way to approach resources sharing and integrated management of distributed component library.In data processing of cloud computing, resources in large number of servers, in different places, are centralized by a major data processing center to manage. In this paper, research on cloud computing applied to software resources management is conducted. In response to the features of managing platform of software resources, software resources, especially management of software resources, especially of component resources are discussed. Hadoop is a source framework based on open source technique and a framework of distributed parallel programming, so distributed programming can be simplified. Hence, to achieve the platform, Hadoop is adopted.Follows are the original points of this thesis:First, the features of cloud computing and relative techniques to achieve it are summarized in this paper. Contrary to the shortage of traditional management of software resources, the research and design of a services platform framework based on clouding computing are conducted.Second, according to the features of managing platform of software resources, Hadoop framework based on cloud computing is adopted to achieve MapReduce parallel programming. Consequently, the cost of development reduced and the efficiency of management and reutilization of software resources increased sharply. Finally, according to the relation between distributed storage of software resources and distributed processing of the platform, a facet is designed, which is applied to retrieve method called " faceted retrieve". The metamessage of the facet is descripted by XML, and a table design is carried out to write the resource information in HBase. To increase the retrieve efficiency of distributed storage resources, the code of offline gathering of search engine on software resources is designed and illustrated too.