云计算环境下服务组合的自治愈关键技术研究

项目名称： 云计算环境下服务组合的自治愈关键技术研究

项目编号： No.61202136

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 计算机科学学科

项目作者： 李雯睿

作者单位： 南京晓庄学院

项目金额： 24万元

中文摘要： 云服务本身的自治性和动态性及环境的异构性等特点引发的多种不确定因素影响了云服务组合的正确性、可靠性和可用性。如何提供持续满足用户需求的云服务组合是当前亟待解决的重大问题。本项目给出了云计算环境下服务组合自治愈技术的解决方案，主要包括三个方面。首先，提出一种基于场景的需求规约语言的扩展形式，来清晰地表示云计算环境下用户的非功能需求，从而提供了一个完全图形化的前端。其次，提出了一种自适应的云服务组合监控技术，分别对基础设施层、平台层和软件层的不同服务资源进行监控，并使得监控机制与被监控云服务组合之间是非纠缠的，以保证监控机制对系统的影响尽可能小。最后，当检测到云服务组合中的失效时，提出一组恢复策略的自动生成技术，并选择最优方案，使恢复策略在实践中是可行的。本研究将进一步揭示云计算环境下服务组合自治愈的一般规律，并为用户提供满足和可靠的云服务提供新方案。

中文关键词： 云计算；场景规约；自治愈；在线监控；恢复

英文摘要： The autonomy and dynamic nature of the cloud services and the heterogeneity of environment may cause a variety of uncertainties. These factors can affect the correctness, reliability and availability of cloud service composition. How to make cloud service composition meet users' requirements is a major issue to be solved. This project gives an innovatiove solution based on self-healing service composition in cloud service environment, including the following three aspects. First, we propose an extension of a scenario-based requirement specification language, to clearly represent the non-functional requirements of the user, which provides a full graphical front end. Secondly, an adaptive cloud service composition monitoring method is proposed to monitor the service resources in the infrastructure layer, platform layer and software layer. It makes the monitoring mechanism and cloud service composition non-intrusive, ensuring that the monitoring mechanism affects the system as small as possible. Finally, when the failures in the cloud services composition are detected, we propose a technique that can automatically generate a set of recovery strategies, and choose the best strategy to make the recovery strategy feasible in practice. This study will further reveal the general laws of self-healing service composition

英文关键词： Cloud computing；scenario-based specifcation；self-healing；online monitoring；recovery