摘要
随着网络的发展，多层分布式集群体系逐渐成为应用趋势，不仅具有更高性能／价格比，而且还提供了良好的可扩展性。在过去，高可用性一般就是需要系统能够提供不间断的服务，系统能够及时的处理硬件故障就可以了。但是，由于全球经济的不断发展，企业对系统的高可用提出了更高的要求。要求系统不但能提供不间断的服务，还要求系统提高在资源管理方面的有效性，不断加强系统在管理上的灵活性。如今在高可用性的基础上自律计算的概念被提了出来。通过自律计算，机器可以自我管理、自我恢复、自我保护、自我优化。换句话说，机器能够像人的神经系统一样工作，在不需要外界的干涉的情况下，控制自己的各项功能。这样的系统能够自我运行，并根据外界的条件及时纠正自己的行为，在各种负载情况下最有效的工作。
本文作者参加了一个高可用系统—虚拟操作环境系统的开发工作，该系统主要用到集群技术，为用户提供了图形化的资源管理工具，实现系统的高可用性和自律计算。在开发过程中，作者对资源的动态分配作了深入的研究。在实现方面，作者就自己参与的策略管理模块，进行了详细的阐述。最后，对系统中存在的问题和后续工作进行了简单的总结。完成了本论文。
虚拟操作环境系统的开发基于可分区服务器灵活的资源配置机制。结合集群管理软件控制服务器，管理共享硬件资源的分区。通过虚拟操作坏境系统，服务器之间可以共享硬件资源，并根据需要重新分配或添加新的硬件，使系统资源得到了高效的利用并使系统具有高可扩展性；当硬件发生故障时，虚拟操作环境系统可提供灵活、及时的系统恢复，提高了系统的可用性。
本文首先对分区和集群技术进行了深入的研究，然后根据多层分布式集群体系的思想，在此基础上设计并实现了虚拟操作环境系统，对虚拟操作环境系统的体系结构，功能模块，理论基础，设计和实现进行了阐述，主要研究了系统的设计和实现。最后，对系统中存在的问题和后续工作进行了简单的总结。
虚拟操作环境系统以随可分区服务器一起投入市场，获得了用户的好评，并取得了良好的经济效益。
With the communication and computer technology developing rapidly , the enterprise more and more depend on the computer system and the network technology. Since the enterprises all over the world think much of the combination of competition and cooperation , there is more and more request for the system High Availability.In the past , the High Availability needs the system provide the continuable service and diSPose the system hardware fault in time. As the world economy developing and the competition picking up , High Availability has the new meaning . It means not only the continuable service, but also the high resource manage efficiency and the convenience facility system management. But now 'autonomic computing' has been presented based of High Availability.The aims of 'autonomic computing' are Self Managing,Self control, Self optimize&Self recovery.
The author of this article be a member of the designing and implementing the HA system Virtual Operation Environment, which based on the cluster, policy based and autonomic computing, providing graphic tools to manager the HA. In the process of development, the author deeply studied the system, and then finished the paper.
Firstly, the paper explains High Availablility and partitionable server , then described the structure and the foundation modules of Virtual Operation Environment system .analyses technology conditions to realize Virtual Operation Environment system, and mainly studies design and realization of the system. In the realization of the system, the author described the database module , policy module and the composition change module in detail, which mainly developed by
herself. At last, analyses the deficiency of the system and the work in the future.
引文
[1] David M．Geary著 李建森 蒋欣军 龚尧莞等译，《Java2图形设计卷Ⅱ：SWING》，北京，机械工业出版社，2000
[2] 刘方生等，《数据库系统概论》，武汉，华中理工大学出版社，1991
[3] Bruce Eckel，《Think in JAVA》，北京，机械工业出版社，1999
[4] 王勇领，《系统分析与设计》，北京，清华大学出版社，1990
[5] Cay S.Horstmann，《Core JAVA2》，北京，机械工业出版社，2000
[6] R. Rajanetal," A policy framework Fro integrated and differentiated Services in the Internet", IEEE NetWork, Vol.13 No. 5, Sepetember/October 1999
[7] Marc H. Willebeek-LeMair, Strategies For Dynamic Load Balancing on Highly Parallel Computers. IEEE Transctions on parallel and distributed systems, VOL 4, No. 9, September 1993
[8] HP System 分区 s Guide Administration for n分区 s Second Edition
[9] 阎宏著《Java与模式》，电子工业出版社，2003
[10] the White Paper of Policy based System.http://www.allot. com
[12] Mary Shaw、David Garlan, 《Software Architecture》,Prentice-Hall International, Inc. 1998
[13] Elias M. Award. menlo Park, 《Concepts, Structure and Apllication》, Calif, Benjamin—Cummings Pubs. Co, 1994
