中国科学院软件研究所机构知识库
Advanced  
ISCAS OpenIR  > 中科院软件所  > 中科院软件所
题名:
Web应用服务器资源管理关键技术研究
作者: 张文博
答辩日期: 2007-01-19
授予单位: 中国科学院软件研究所
授予地点: 软件研究所
学位: 博士
关键词: Web应用服务器 ; 资源管理 ; 性能 ; 服务质量 ; 资源重配 ; 调度资源管理 ; 中间件
其他题名: Research on Key Technologies in Resource Management of Web Application Server
摘要: Web应用服务器已经成为面向Web计算环境的主流中间件平台,它为创建、部署、运行、集成和管理企业级应用提供了基础设施支持。随着企业级应用非功能性需求的不断提高,Web应用服务器的性能和QoS(Quality of Service)保障能力受到越来越多的关注,而提高这些能力的关键之一是Web应用服务器的资源管理技术。 资源管理技术虽然在一般的中间件领域已有大量的研究成果,但是Web应用服务器面向的是高度动态变化的Web计算环境,大量的不可预知因素(运行环境、负载、客户访问模式…)对资源管理技术提出了新的挑战:Web应用服务器固定的资源配置难以满足应用动态变化的资源需求,因而必须能在运行时灵活地重配资源以缓解由此而引起的性能问题;在同时支持多个应用时,现有的服务器难以在应用间高效地协调资源利用;Web计算环境固有的过载会危害Web应用服务器提供给应用的QoS保障等。目前已有的相关研究在资源重配效率以及稳定性方面还存在不足,对应用的QoS需求特征分析不够充分,固化的调度资源管理机制难以充分满足应用灵活的QoS需求。本文对这些问题进行了针对性研究。 首先,研究了Web应用服务器自适应资源重配技术。针对Web应用服务器的资源特点,提出了一种包含物理资源和逻辑资源在内的两级资源结构,并基于该结构给出资源重配目标与重配规则,设计了高效的自适应资源重配机制。该机制使用较为稳定的资源状态(相比于性能指标)作为资源重配目标,实现了高效的资源重配执行机制。与相关工作相比,它使Web应用服务器在更短的时间内获得比较理想的资源配置,并且不易发生资源配置的抖动,从而提高了Web应用服务器资源重配的效率以及对动态计算环境的适应能力。 在支持多个应用同时运行方面,研究了一种基于域的多应用资源重配机制,与已有工作隔离应用资源空间的方法不同,本文采取一种共享的多应用资源分配模式,并设计基于资源报价的多应用资源重配算法以解决应用间的资源需求冲突。实验结果表明它可以根据应用的实际的资源需求进行应用间的资源重配,并提高了Web应用服务器整体的性能和资源在应用间的利用效率。 通过分析Web应用服务器典型应用场景(Web站点)的QoS需求特点,提出了一种针对调度资源管理的QoS需求描述方法。该方法增加了对影响请求重要性的动态因素的描述方式,并根据客户的QoS感受变化特点提出了一种可降级的请求时间约束描述方式,这些特点增强了对QoS需求业务特征的表达能力。通过使用QoS需求中包含的请求调度信息,本文研究了QoS需求驱动的调度资源管理技术,包括针对可降级时间约束需求的负载管理技术以及基于获益风险的调度资源管理策略,并在此基础上设计了一种QoS需求驱动的调度资源管理机制。其特点在于通过充分使用QoS需求中提供的请求调度信息,满足应用对请求时间约束的需求并大幅提高对应用业务目标的保障能力,同时可以提供优雅降级的QoS保障。 最后,以上资源管理技术研究成果在一个高性能的Web应用服务器系统OnceAS2.0中进行了应用。在J2EE性能测试基准ECperf的测试中,OnceAS2.0表现出良好的性能和QoS保障能力。
英文摘要: Web Application Server (or WAS for short) has become one of mainstream middleware technologies for the Web computing environments. It provides infrastructural support for fast development, deployment, running, integration, and management of enterprise-level applications. With the increasing requirements on the non-functional properties of enterprise-level applications, WAS’s performance and Quality of Service (or QoS for short) have received more and more attention. Both the performance and QoS issues relate to resource management technology that is the focus of this thesis. Generally, resource management technology has been studied extensively by the middleware community. Current progress of resource management technology for general middleware, however, provides inadequate support for WAS. The main reason is that the Web computing environments targeted by WAS are highly dynamic. For example, the runtime environments, workloads, and user patterns of WAS applications typically change in an unpredictable manner. This fact makes difficult for WAS to provide sufficient performance and QoS guarantees for its applications. First, static resource management in a WAS can easily fail to meet dynamic resource requirements from applications. Hence, WAS needs to make its resource management adaptable to actual requirements on resources from applications by automatic reconfiguration. Second, WAS is difficult to share resource between applications efficiently. Third, overload is one of inherent features of WAS computing and hence a successful WAS needs to provide sufficient QoS guarantee to handle overload through suitable scheduling in its resource management. However, WAS has no thorough analysis of QoS requirements and corresponding scheduling resource management mechanism, thus fail to satisfy the flexible QoS requirements of applications. First, we propose a resource adaptation technique to tackle changeable, unpredictable resource requirements imposed by dynamic Web computing environments. We present a two-layered architecture for the resource management in WAS. This architecture concerns both physical resources and logical resources. Reconfiguration rules are studied for better resource management by adapting resource allocation according to dynamic environments. A high efficient adaptive reconfiguration mechanism is then presented based on the above technologies. We adopt the resource state as the reconfiguration target, which is more stable than other metrics, and design a fast reconfiguration approach. Compared with the existing work, WAS can achieve favorable configuration and behave more stablely, which indicate that the improvement of reconfiguration efficiency and adaptability. Second, based on the above resource adaptation technique, we extend it to support realm-based multiple applications with resource reconfiguration capabilities. This extension consists of a resource allocation model to handle shared resources among applications and a resource bidding algorithm to resolve resource conflicts at runtime. The experiment results show that the whole performance of WAS is raised by applying the mechanism, which indicate that the resources have been highly reused between applications. Third, we study how to support QoS requirements from applications directly. We put our study in typical scenarios of WAS applications, and propose a requirement declaration mechanism for applications to claim their QoS requirements. This mechanism considers dynamic factors such as the change of request importance and deadline of request processing. Such factors play an important role in our QoS-driven schedule model to manage resource allocation for applications, and offer better QoS guarantees. According to the experiment results, the business metrics and deadlines are satisfied better and graceful degrade is also provided. Finally, we have integrated all the above techniques together and implemented them in our WAS system - OnceAS2.0. From our experiments based on benchmark ECperf, OnceAS2.0 has satisfactory performance and QoS guarantees.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.iscas.ac.cn/handle/311060/6668
Appears in Collections:中科院软件所

Files in This Item:
File Name/ File Size Content Type Version Access License
10001_200218015003317张文博_paper.pdf(2597KB)----限制开放-- 联系获取全文

Recommended Citation:
张文博. Web应用服务器资源管理关键技术研究[D]. 软件研究所. 中国科学院软件研究所. 2007-01-19.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[张文博]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[张文博]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2017  中国科学院软件研究所 - Feedback
Powered by CSpace