云环境下基于QoS和负载平衡的服务选择方法研究

发布时间：2018-05-22 20:21

本文选题：Web服务 + 服务质量　；参考：《湖南科技大学》2014年硕士论文

【摘要】：在云环境下，越来越多的功能相同或相似但非功能属性不同的Web服务被采用。服务质量（Quality of Services, QoS），成为了Web服务选择中服务提供者赢得市场的关键非功能属性之一。随着Internet技术的日益发展，用户能够方便地按需获取满足其要求的服务资源，而高质量的服务将受到用户的青睐。然而，云环境下的服务请求具有即时、并发以及大规模等特点，多个用户可能同时调用同一个高质量的服务，极有可能出现用户访问量超过服务的负载容量的情况，致使服务负载失衡，从而导致无法即时响应用户请求，服务提供能力急剧下降。针对该问题，近年来已有少量学者考虑了服务的负载对服务性能的影响，并且在这方面已经展开了一些研究。然而，目前仍存在着很多亟待解决的问题：当多个用户同时访问单个服务或多个服务时，如何合理的为用户提供满意的服务并保证服务负载平衡；当功能单一或数量有限的服务不能满足用户需求时，如何组合这些服务并考虑服务负载的影响以高质量地为用户提供服务。针对这些问题，本文开展了如下工作： (1)提出了一种基于QoS和负载平衡的单服务选择方法。目前的服务选择方法大多是从用户角度出发，从众多服务中选择高质量的服务给用户。然而，云环境下，随着Web服务的用户访问量增加，将出现多个用户同时访问单个服务的情况，，服务负载能力将急剧下降，服务的响应时间将延长，致使用户选择不到高质量的服务。因此，负载感知的服务选择方法能在保证负载平衡的情况下有效地帮助用户选择满足需求的高质量的服务。 (2)提出了一种云环境下基于QoS效应值和负载平衡的全局优化多服务选择方法。在云环境下，同一时刻或极短的时间内可能存在多个服务提供者和请求者分别正在提供和请求服务。在这种情况下，不仅要提供最满意的服务给用户，也要确保整个服务系统的负载平衡。因此，应在同一时刻或在极短的时间内为多个用户提供满意的服务，从而使得服务的用户满意度整体最大，同时使得服务系统的负载整体最小。 (3)在服务组合过程中，引入服务负载参数，设计基于QoS与服务负载的服务组合模型及算法，并利用粒子群算法进行优化。在云环境下，用户在选择服务的过程中，满足其需求的不一定是单个服务，也极有可能是组合服务。目前的服务组合方法，充分重视了服务QoS（如：执行价格、响应时间等）对服务组合的影响，国内外学者也提出了很多的基于QoS的服务组合及优化方法，但它们大都未考虑服务负载变化对服务组合性能的影响，这将可能导致满足用户需求的服务组合失效。因此，该方法不仅能够为用户提供高质量的组合服务，而且保障了组合服务的负载平衡，大大提高了服务组合的成功率。
[Abstract]:In the cloud environment, more and more Web services with the same or similar functions but different non-functional attributes are adopted. Quality of Service (QoS) has become one of the key non-functional attributes for service providers to win the market in Web service selection. With the development of Internet technology, users can easily get service resources to meet their needs on demand, and high quality services will be favored by users. However, the service request in the cloud environment has the characteristics of instant, concurrency and large scale. Multiple users may call the same high quality service at the same time, and it is very likely that the number of users visiting the service exceeds the load capacity of the service. As a result, the service load is out of balance, which leads to the failure to respond to the user's request immediately, and the service delivery capacity drops sharply. In recent years, a small number of scholars have considered the impact of service load on service performance, and some studies have been carried out in this area. However, there are still many problems to be solved: how to provide satisfactory services and ensure load balance for users when multiple users visit a single service or multiple services at the same time; When a single or limited number of services can not meet the needs of users, how to combine these services and consider the impact of service load to provide services to users with high quality. In view of these problems, this paper has carried out the following work: A single service selection method based on QoS and load balancing is proposed. At present, most of the service selection methods are from the user's point of view, and select the high quality service to the user from the numerous services. However, in the cloud environment, as the number of users of Web services increases, there will be multiple users accessing a single service at the same time, the load capacity of the service will decline dramatically, and the response time of the service will be prolonged. Causes the user to choose the high quality service. Therefore, the load-aware service selection method can effectively help users to select high-quality services to meet their needs while ensuring load balance. A global optimal multi-service selection method based on QoS effect and load balancing in cloud environment is proposed. In a cloud environment, multiple service providers and requesters may be providing and requesting services at the same time or in a very short period of time. In this case, not only to provide the most satisfactory service to users, but also to ensure the load balance of the entire service system. Therefore, it is necessary to provide satisfactory services to multiple users at the same time or in a very short time, so as to maximize the overall customer satisfaction of the service, and at the same time to minimize the overall load of the service system. In the process of service composition, service load parameters are introduced, service composition model and algorithm based on QoS and service load are designed, and particle swarm optimization (PSO) is used to optimize the model. In the cloud environment, in the process of choosing services, the users may not satisfy the needs of a single service, but also may be a composite service. In the current service composition methods, the influence of QoS (such as execution price, response time, etc.) on service composition has been fully taken into account. Many service composition and optimization methods based on QoS have also been proposed by domestic and foreign scholars. However, most of them do not consider the impact of service load changes on service composition performance, which may lead to the failure of service composition to meet the needs of users. Therefore, this method can not only provide high quality composite services for users, but also ensure the load balance of composite services, and greatly improve the success rate of service composition.
【学位授予单位】：湖南科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】