计算机科学与探索 ›› 2015, Vol. 9 ›› Issue (6): 669-681.DOI: 10.3778/j.issn.1673-9418.1410006

• 系统软件与软件工程 • 上一篇    下一篇

面向用户体验的Web软件系统优化控制方法

俞  岭1,2+,树  岸1,2,吴毅坚1,2,彭  鑫1,2,赵文耘1,2   

  1. 1. 复旦大学 软件学院,上海 201203
    2. 复旦大学 上海市数据科学重点实验室,上海 201203
  • 出版日期:2015-06-01 发布日期:2015-06-04

Optimal Control Approach of Web Software Systems Based on User Experience

YU Ling1,2+, SHU An1,2, WU Yijian1,2, PENG Xin1,2, ZHAO Wenyun1,2   

  1. 1. Software School, Fudan University, Shanghai 201203, China
    2. Shanghai Key Laboratory of Data Science, Fudan University, Shanghai 201203, China
  • Online:2015-06-01 Published:2015-06-04

摘要: 在环境动态变化和系统状态无法预知的情况下,如何同时满足功能性和非功能性需求是一个极具挑战的问题。传统的方法是利用请求调度策略对Web服务请求进行优化,以解决应用的超负荷问题, 但这种方法并不能完全满足QoS复杂性和灵活性的需要。基于系统管理员设置的自适应方法在很大程度上依赖于管理员的主观判断和实践经验,很多情况下人为的预估结果与实际效果存在较大差距。针对上述问题,提出了一个基于用户体验的动态QoS优化方法。该方法使用一个用户体验计算模块动态计算系统响应时间的期望值,一个外层反馈控制器动态调整虚拟资源的个数,并结合一个内层反馈控制器来动态调整各个虚拟资源处理的负载比例关系,从而实现了系统的动态QoS优化。通过与简单规则调节动态QoS的方法进行对比实验,验证了所提方法的有效性。

关键词: 用户体验, 动态QoS优化方法, 反馈控制器, 虚拟资源

Abstract: In the case of dynamic environment and unpredictable system status, how to satisfy both functional and non-functional requirements is a very challenging problem. The traditional approach dealing with overload issue is to optimize Web service request by using request scheduling strategy. But it cannot meet the requirement of complexity and flexibility of QoS completely. Self-adaptation approach based on settings of system administrators largely depends on their subjective judgments and practical experiences. In many cases, there is a large gap between people?s estimates and actual effects. In response to these problems, this paper presents a dynamic QoS optimization approach based on user experience. The approach realizes system dynamic QoS optimization by employing a dynamic user experience computing module to compute expected response time, an outer feedback controller to adjust the number of virtual resources and an inner feedback controller to adjust the weight of each virtual resource. By comparison with a simple-rule method, this paper verifies the effectiveness of the proposed approach.

Key words: user experience, dynamic QoS optimization approach, feedback controller, virtual resource