云计算环境下WEB应用的自动部署与资源分配

发布时间：2018-05-17 11:59

本文选题：云计算 + Web应用　；参考：《哈尔滨工业大学》2014年硕士论文

【摘要】：在云计算环境下云服务提供商管理着大量的Web应用，又以购买或租用的方式获取基础设施，通过静态部署与动态资源分配以Web应用的形式为租户提供软件服务。通过合理的对Web应用进行部署，云服务提供商可以有针对性的提高资源的有效利用率，或降低基础设施的运行成本；根据Web应用的负载与节点负载动态的为租户的Web应用请求进行资源分配，，可以实现基础设施的负载均衡，保证租户服务请求的服务质量，达到资源的合理利用。然而，在云计算环境下，Web应用的组织方式更加灵活，调度与管理的粒度更加精细。许多耦合度低，功能划分明确的Web应用被划分为多个应用构件，其部署与资源分配的单位由原来的应用实例变为构件实例。Web应用的服务模型与架构发生了变化，使传统的Web应用部署与资源分配方式对Web应用管理效率产生了制约，需要新的部署与资源分配方式。针对以上问题，本文主要进行以下工作：（1）针对云计算环境下Web应用的描述问题，提出云计算环境下多Web应用的服务架构，又在此架构基础上给出构件化Web应用服务模型，对模型中的元素与元素的映射关系进行描述，将云计算环境下Web应用的部署与资源分配问题归结为构件化Web应用的部署与资源分配问题，并从概念、资源需求、资源计算方法等方面对Web应用构件进行详细描述。（2）针对构件化Web应用的部署问题，对构件化Web应用部署问题进行形式化描述，设计并实现基于混合遗传模拟退火算法的构件化Web应用部署算法，通过实验对影响部署效率的算法参数进行优化，并设计模拟实验，通过与其他启发式算法进行实验对比，验证算法的正确性与有效性。（3）针对构件化Web应用的资源分配问题，将构件化Web应用的资源分配归结为负载均衡问题，对构件化Web应用负载均衡问题进行分析，定义构件化Web应用负载均衡参数，设计并实现SLA驱动的构件化Web应用动态负载算法，设计对比模拟实验，对实验结果进行分析。（4）针对Web应用的自动部署问题，本文分析云计算架构，给出云计算环境下云服务提供商对Web应用的管理需求，设计现构件化Web应用自动部署工具，实现对构件化Web应用的自动部署。
[Abstract]:In the cloud computing environment, cloud service providers manage a large number of Web applications, acquire infrastructure by buying or renting, and provide software services to tenants through static deployment and dynamic resource allocation in the form of Web applications. Through reasonable deployment of Web applications, cloud service providers can improve the effective utilization of resources or reduce the running cost of infrastructure. According to the load of Web application and the load of node, the resource is allocated dynamically for the tenant's Web application request, which can realize the load balance of infrastructure, guarantee the service quality of tenant service request, and achieve the reasonable utilization of resources. However, in cloud computing environment, the organization of Web applications is more flexible, and the granularity of scheduling and management is more fine. Many Web applications with low coupling degree and clear function division are divided into multiple application components, and the units of deployment and resource allocation have changed from the original application instance to the service model and architecture of the component instance .web application. The traditional way of Web application deployment and resource allocation has restricted the efficiency of Web application management, and a new way of deployment and resource allocation is needed. In view of the above problems, the main work of this paper is as follows: 1) aiming at the description of Web application in cloud computing environment, the service architecture of multi-Web application in cloud computing environment is proposed, and the component-based Web application service model is given based on this architecture, and the mapping relationship between elements and elements in the model is described. The deployment and resource allocation of Web applications in cloud computing environment is reduced to the deployment and resource allocation problem of component-based Web applications. The concept, resource requirements and resource calculation methods of Web application components are described in detail. 2) aiming at the deployment of component-based Web applications, the deployment problem of component-based Web applications is formally described, and a component-based Web application deployment algorithm based on hybrid genetic simulated annealing algorithm is designed and implemented. The algorithm parameters which affect the efficiency of deployment are optimized through experiments. Simulation experiments are designed and compared with other heuristic algorithms to verify the correctness and effectiveness of the algorithm. 3) aiming at the resource allocation problem of component Web application, the resource allocation of component Web application is reduced to load balancing problem, and the load balancing problem of component Web application is analyzed, and the load balancing parameters of component Web application are defined. The dynamic load algorithm of component based Web driven by SLA is designed and implemented. The simulation experiment is designed and compared, and the experimental results are analyzed. 4) aiming at the problem of automatic deployment of Web applications, this paper analyzes the cloud computing architecture, presents the management requirements of cloud service providers for Web applications in cloud computing environment, and designs a component-based automatic deployment tool for Web applications. The automatic deployment of component based Web application is realized.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】