基于服务质量的云工作流调度工具的研究与应用

发布时间：2018-09-07 21:19

【摘要】：针对工作流系统中的不能很好满足用户的多服务质量需求和不能高效调度多样化资源的问题,本课题在云工作流调度工具的体系结构及CloudSim模型扩展、基于服务质量的云工作流调度算法、工具应用等方面开展研究,具体归结为以下几个方面:1、云工作流调度工具软件体系结构及CloudSim模型扩展研究:阐述了云工作流调度工具本身具有的特点及应有的功能。在可信工作流管理系统参考模型的基础上,结合云环境下资源部署结构,增加用户接口、QoS管理、云资源管理等,提出一种云工作流调度工具软件体系结构,同时根据工作流特点,对已存在的QoS结构进行改进,增加QoS预测及反馈功能。在云工作流调度工具体系结构的基础上,以CloudSim为平台,扩展了平台的工作流调度功能和工作流调度算法,形成了基于CloudSim扩展的云工作流调度工具。2、基于服务质量的云工作流调度算法研究:从云工作流调度工具中的云工作流调度模块出发,改进了蚁群算法,引入云任务优先次序启发式规则,形成了基于服务质量的云工作流调度算法。同时,通过科学工作流和实例密集型工作流对云工作流调度算法进行测试和验证,结果表明这两种工作流实例在执行时间和执行成本上表现出较为严格的一致性,即改进的蚁群算法可以满足功能性需求,还能大幅降低执行时间和耗费成本;改进的蚁群算法在不同场景,尤其是在用户具有不同程度的偏好时,有显著优势。3、调度工具设计与应用研究:在CloudSim平台上完成基于工作流的CloudSim模型扩展,实现了基于服务质量的云工作流调度算法。以网购流程为例,阐述了实例密集型工作流在云工作流调度工具中的整个运行过程,实现了云工作流调度算法在实例中的应用。
[Abstract]:In order to solve the problem that the multi-QoS requirements of users can not be well met and the diversified resources can not be efficiently scheduled in workflow system, this paper extends the architecture of cloud workflow scheduling tools and CloudSim model. Research on cloud workflow scheduling algorithm, tool application and so on based on quality of service, This paper summarizes the following aspects: 1, the research on the software architecture of cloud workflow scheduling tools and the extension of CloudSim model: the characteristics and functions of cloud workflow scheduling tools are expounded. Based on the reference model of trusted workflow management system, combined with resource deployment structure under cloud environment, adding user interface to QoS management and cloud resource management, a cloud workflow scheduling tool software architecture is proposed. At the same time, according to the characteristics of workflow, the existing QoS structure is improved to increase the function of QoS prediction and feedback. Based on the architecture of cloud workflow scheduling tool, the workflow scheduling function and workflow scheduling algorithm of the platform are extended based on CloudSim. Cloud workflow scheduling tool based on CloudSim extension. 2. Research on cloud workflow scheduling algorithm based on quality of service. Based on cloud workflow scheduling module in cloud workflow scheduling tool, ant colony algorithm is improved. A cloud workflow scheduling algorithm based on quality of service (QoS) is proposed by introducing the cloud task priority heuristic rule. At the same time, the cloud workflow scheduling algorithm is tested and verified by scientific workflow and instance intensive workflow. The results show that these two workflow instances have strict consistency in execution time and execution cost. That is, the improved ant colony algorithm can satisfy the functional requirements and greatly reduce the execution time and cost. The improved ant colony algorithm can be used in different scenarios, especially when users have different degrees of preference. There are significant advantages. 3. Research on the design and application of scheduling tools: the CloudSim model extension based on workflow is completed on CloudSim platform, and the cloud workflow scheduling algorithm based on quality of service is implemented. Taking the online shopping process as an example, the whole running process of instance intensive workflow in cloud workflow scheduling tool is expounded, and the application of cloud workflow scheduling algorithm in instance is realized.
【学位授予单位】：北方民族大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52

【参考文献】