云移植中遗留系统组件重用研究

项目名称： 云移植中遗留系统组件重用研究

项目编号： No.61462066

项目类型： 地区科学基金项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 赵俊峰

作者单位： 内蒙古大学

项目金额： 45万元

中文摘要： 云计算提供易重用、可扩展的虚拟资源，这些资源可以被动态配置以实现良好的扩展性和优化的资源使用。为了利用云计算的优势和保护对遗留软件的已有投资，许多组织希望能够将遗留软件移植到云平台。本项目针对云移植过程中遗留系统的组件重用方法开展研究，主要包括以下三部分内容。（1）正交体系结构恢复方法，该方法结合垂直聚类和水平聚类，垂直聚类实现系统功能分解，水平聚类实现体系结构分层，由此准确发现可重用逻辑组件并生成与之对应的组件描述。（2）服务发现，通过分析业务需求识别目标服务，设计匹配算法实现可重用组件与目标服务的映射。（3）服务实现，依据云平台编程模型对可重用组件进行算法重构，从而有效利用云平台的资源。本项目提出的方法将使云移植中遗留软件得到有效重用，提高软件移植效率，更好保证目标服务的质量。

中文关键词： 云计算；软件即服务；体系结构恢复；可重用组件；聚类

英文摘要： Cloud computing can provide reusable,scalable virtual resources,which can be configured dynamically to realize good extensibility and optimal resource usage. For the purpose of taking advantage of cloud computing and protecting the assets of legacy software, most organizations expect to migrate the legacy software system to cloud platform. This project tries to carry out research on the approach on component reuse of legacy system in cloud migration, which includes three main aspects as follows.(1) An approach of orthogonal architecture recovery integrates vertical clustering and horizontal clustering. Vetical clustering is used to decompose legacy system into function modules, and horizontal clustering realizes structure layering, through which the reusable logic components can be discovered precisely and the corresponding description information be generated.(2) Service discovery identifies target services by analyzing the business requirement, and maps the reusable components and target services by designing a related match algorithm. (3) Service implementation refactors the algorithms in components according to the programming models in cloud platform, so the advantage of the cloud can be utilized efficiently. The approach in this project will make the implementation of legacy system be reused efficiently, improve the efficiency of cloud migration and guarantee the quality of target services.

英文关键词： Cloud Computing;Software as a Service;Architecture Recovery;Reusable Component;Clustering