基于Web Service的分布式系统中动态负载均衡策略的研究与实现
|
摘要
目前,分布式系统及其分布式计算技术是计算机领域的热点问题。在对分布式系统的研究中人们注意到,在某一时刻,一些计算机系统的负载很轻,而另外一些计算机系统的负载却很重,整个系统的效率并不高,所以,采用有效的负载均衡策略就成为了提高分布式系统性能的主要手段。虽然负载均衡问题至今已有20多年的研究历史,但由于负载均衡是NP完全问题,因此,目前还没有一种有效的负载均衡策略能够适用于所有的环境,负载均衡问题还有待于进一步的研究。负载均衡可分为静态负载均衡和动态负载均衡两类。实验表明,通常情况下动态负载均衡比静态负载均衡的性能效果有30%～40%的提高,因此本文研究的是动态负载均衡策略,具体的研究内容如下:
首先,根据实际需求,设计了一个基于Web Service的分布式决策支持系统。为了解决以往C/S、B/S结构的诸多问题,本文采用了B/S与Web服务相结合的整体结构,并采用Web Service技术实现信息的传递,实现了异构平台间跨越防火墙的互操作功能,同时也增强了系统的可扩展性。
其次,基于提出的决策支持系统,结合已有的动态负载均衡策略,设计了一个适用于该系统的改进的动态负载均衡策略。该策略采用的是全局的、集中式的、周期性的启动策略,设置了一个管理者周期性的收集系统内各节点的状态信息,并能够根据系统状态的变化,决定由谁来触发负载分配,这样使得该策略在任何状态下都能保证合理的负载均衡活动。
最后,实现设计的系统及动态负载均衡策略,并通过对实验结果进行分析,验证该策略的稳定性和有效性。
Currently, distributed system and distributed computing technology is a hot issue of the computer field. In many studies of distributed system it was noted that, at some point, the load of some computer systems is very light, while the load of others is very heavy, the efficiency of whole system is not high, therefore, effective load balancing strategy has become the primary way to improve the performance of distributed system. Load balancing has been researched for 20 years, but because it is a NP problem, there is no an effective load balancing strategy can be applied to all environments, this problem needs to be further studied. Load balancing can be divided into static load balancing and dynamic load balancing. Studies have shown that compared with static load balancing, the performance of dynamic load balancing can usually increase 30% to 40%, therefore, dynamic load balancing strategy is the focus of the paper, the specific contents are as follows:
Firstly, according to the actual needs, a distributed decision support system based on Web Service is designed. In order to solve previous problems in C/S and B/S structure, B/S and Web service are applied to design the structure of this system, and Web Service technology is adopted to transfer the information and achieve interoperability functions across the firewall in heterogeneous platforms, the scalability of the system is also enhanced.
Secondly, combined with existing dynamic load balancing strategy, it proposes an improved dynamic load balancing strategy based on this system. The strategy adopts an overall, central controlled, periodic start strategy, and sets up a manager to collect the status information of all nodes in this system periodically, and according to the change of system state, it decides that who is going to trigger the load distribution, so this strategy can ensure a reasonable load balancing activity in any condition.
Finally, the system and the dynamic load balancing strategy have been implemented, and the experimental results are analyzed in order to verify the stability and effectiveness of the strategy.
首先,根据实际需求,设计了一个基于Web Service的分布式决策支持系统。为了解决以往C/S、B/S结构的诸多问题,本文采用了B/S与Web服务相结合的整体结构,并采用Web Service技术实现信息的传递,实现了异构平台间跨越防火墙的互操作功能,同时也增强了系统的可扩展性。
其次,基于提出的决策支持系统,结合已有的动态负载均衡策略,设计了一个适用于该系统的改进的动态负载均衡策略。该策略采用的是全局的、集中式的、周期性的启动策略,设置了一个管理者周期性的收集系统内各节点的状态信息,并能够根据系统状态的变化,决定由谁来触发负载分配,这样使得该策略在任何状态下都能保证合理的负载均衡活动。
最后,实现设计的系统及动态负载均衡策略,并通过对实验结果进行分析,验证该策略的稳定性和有效性。
Currently, distributed system and distributed computing technology is a hot issue of the computer field. In many studies of distributed system it was noted that, at some point, the load of some computer systems is very light, while the load of others is very heavy, the efficiency of whole system is not high, therefore, effective load balancing strategy has become the primary way to improve the performance of distributed system. Load balancing has been researched for 20 years, but because it is a NP problem, there is no an effective load balancing strategy can be applied to all environments, this problem needs to be further studied. Load balancing can be divided into static load balancing and dynamic load balancing. Studies have shown that compared with static load balancing, the performance of dynamic load balancing can usually increase 30% to 40%, therefore, dynamic load balancing strategy is the focus of the paper, the specific contents are as follows:
Firstly, according to the actual needs, a distributed decision support system based on Web Service is designed. In order to solve previous problems in C/S and B/S structure, B/S and Web service are applied to design the structure of this system, and Web Service technology is adopted to transfer the information and achieve interoperability functions across the firewall in heterogeneous platforms, the scalability of the system is also enhanced.
Secondly, combined with existing dynamic load balancing strategy, it proposes an improved dynamic load balancing strategy based on this system. The strategy adopts an overall, central controlled, periodic start strategy, and sets up a manager to collect the status information of all nodes in this system periodically, and according to the change of system state, it decides that who is going to trigger the load distribution, so this strategy can ensure a reasonable load balancing activity in any condition.
Finally, the system and the dynamic load balancing strategy have been implemented, and the experimental results are analyzed in order to verify the stability and effectiveness of the strategy.
引文
[1] H E.Rewini and T.G.Lewis.Task Scheduling in Parallel and Distributed Systems.New Jersey: Prentice Hall, 1994
[2] Lan Z, Taylor V E, Bryan G.Dynamic load balancing for adaptive mesh refinement application.Parallel Processing(ICPP),Northwestern Univ, 2001: 571-579P
[3] Andrew S. Tanenbaum著.分布式操作系统.陆丽娜,伍卫国,刘隆国译.北京:电子工业出版社,2008
[4]何炎祥.分布式操作系统.北京:高等教育出版社,2005
[5] Andrew S,Tanenbaum,Maarten Van Steen著.分布式系统原理与范型.辛春生,陈宗斌译.北京:清华大学出版社,2008
[6] Tanenbaum A S.Distributed Operating Systems.北京:电子工业出版社,1999
[7]樊胜.C/S与B/S的结构比较及Web数据库的访问方式.情报科学.2001,19(4):443-445页
[8]林海.浏览器/服务器应用开发[M].科学出版社,2000
[9] Tsur S.Are Web Services the Next Revolution in E-Commerce?[A].Proc of VLDB 2001[C].2001: 614-617P
[10]郑晓东,王志坚,周晓峰,费玉奎,许峰.一种基于Web Service的分布式计算模型研究及其实现.计算机工程与应用.2004,(1):145-146页
[11]柴晓路,梁宇奇.Web Services技术、架构和应用.电子工业出版社,2003:89-94页
[12] IBM developer works . Web architecture: Web Service architecture overview[R].2000: 3-5P
[13]李淑艳.Web Service技术研究与实现.电脑开发与应用.2008,21(11):42-44页
[14] Florescu D, Grunhagen A, Kossmann D.An XML programming languagefor Web service specification and composition.Proc. of the 11th Int’1 World Wide Web Conf. Honolulu: ACM, 2002: 65-76P
[15]吕曦,王化文.Web Service的架构与协议.计算机应用.2002,22(12):62-65页
[16] E Christensen, F Curbera, G Meredith et.Web Services Description Language(WSDL)1.1 http://www.w3.org/TR/wsdl
[17]杨涛,刘锦德.Web Service技术综述—一种面向服务的分布式计算模式.计算机应用.2004,24(8):1-4页
[18]饶元.Web Services:分布式网络体系新架构.计算机工程.2004,30(22):1-3页
[19]杨德仁,薛梅,顾君忠.Web Services核心协议与实施研究.计算机应用系统.2005,(1):33-36页
[20]王楠,刘心雄,陈和平.Web Service技术研究.计算机与数字工程.2006,34(7):88-90页
[21] Kunz, T..The influence of different workload descriptions on a heuristic load balancing scheme[J].IEEE Transactions on Software Engineering, 1991, 17(7): 725-730P
[22] Kameda H, Fathy S, Ryu I et a1.A performance comparison of dynamic vs. static load balancing policies in a mainframe-personal computer network model.Proceedings of IEEE CDC2000, Sydney, 2000: 1415-1420P
[23]陈涛,陈启买.分布式计算机系统负载平衡研究.计算机技术与发展.2006,16(5):33-35页
[24] Eager, D.L., E.D.Lazowska, and J.Zahorijan . A comparison of receiver-initiated and sender-initiated adaptive load sharing[J].performance Evaluation, 1986, 1(3): 53-68P
[25] Lin, F.C.H. and R.M.Keller . The gradient model load balancing method.IEEE Transactions on Software Engineering, 1987, 13(1): 32-28P
[26] Bryant, R.M. and R.A.Finkel . A stable distributed scheduling algorithm.Proc. of the 2~(nd) int’1 Conf. on Distributed Computing Systems, 1981: 314-323P
[27]李冬梅,施海虎.负载平衡调度问题的一般模型研究.计算机工程与应用.2007,43(8):121-125页
[28] Singhal, M. and N.G. Shivaratri.Advanced Concepts in Operating Systems: Distributed,Database, S. and Multiprocessors Operating Systems, McGraw-Hill, Inc., 1994
[29] Leff, A. and P.S. Yu.An adaptive strategy for load sharing in distributed database environments with information lags.Journal of Parallel and Distributed Computing, 1991, 13(1): 91-103P
[30]彭玲玲.改进的分布式动态自适应负载平衡算法探讨.电脑学习.2009,(3):134-135页
[31]罗拥军,李晓乐,孙如祥.负载均衡算法综述.科技情报开发与经济.2008,18(23):134-135页
[32]王友良,叶柏龙.分布式系统中动态负载平衡的研究.科学技术与工程.2005,5(9):572-574页
[33]李岱松,刘勇,潘成胜.一种负载均衡算法的研究与实现.沈阳理工大学学报.2005,24(3):47-48页
[34]张宇晴,佟振声,胡旦华.分布式系统中动态负载平衡算法的研究.计算机仿真.2003,20(9):69-70页
[35] Suen, T. T. Y. and J. S. K. Wong.Efficient task migration algorithm for distributed systems.IEEE Transactions on Parallel and Distributed Systems, 1992, 3(4): 488-499P
[36]刘建,李绪志.一种动态负载均衡机制的研究与实现.计算机工程与应用.2006,(2):142-144页
[2] Lan Z, Taylor V E, Bryan G.Dynamic load balancing for adaptive mesh refinement application.Parallel Processing(ICPP),Northwestern Univ, 2001: 571-579P
[3] Andrew S. Tanenbaum著.分布式操作系统.陆丽娜,伍卫国,刘隆国译.北京:电子工业出版社,2008
[4]何炎祥.分布式操作系统.北京:高等教育出版社,2005
[5] Andrew S,Tanenbaum,Maarten Van Steen著.分布式系统原理与范型.辛春生,陈宗斌译.北京:清华大学出版社,2008
[6] Tanenbaum A S.Distributed Operating Systems.北京:电子工业出版社,1999
[7]樊胜.C/S与B/S的结构比较及Web数据库的访问方式.情报科学.2001,19(4):443-445页
[8]林海.浏览器/服务器应用开发[M].科学出版社,2000
[9] Tsur S.Are Web Services the Next Revolution in E-Commerce?[A].Proc of VLDB 2001[C].2001: 614-617P
[10]郑晓东,王志坚,周晓峰,费玉奎,许峰.一种基于Web Service的分布式计算模型研究及其实现.计算机工程与应用.2004,(1):145-146页
[11]柴晓路,梁宇奇.Web Services技术、架构和应用.电子工业出版社,2003:89-94页
[12] IBM developer works . Web architecture: Web Service architecture overview[R].2000: 3-5P
[13]李淑艳.Web Service技术研究与实现.电脑开发与应用.2008,21(11):42-44页
[14] Florescu D, Grunhagen A, Kossmann D.An XML programming languagefor Web service specification and composition.Proc. of the 11th Int’1 World Wide Web Conf. Honolulu: ACM, 2002: 65-76P
[15]吕曦,王化文.Web Service的架构与协议.计算机应用.2002,22(12):62-65页
[16] E Christensen, F Curbera, G Meredith et.Web Services Description Language(WSDL)1.1 http://www.w3.org/TR/wsdl
[17]杨涛,刘锦德.Web Service技术综述—一种面向服务的分布式计算模式.计算机应用.2004,24(8):1-4页
[18]饶元.Web Services:分布式网络体系新架构.计算机工程.2004,30(22):1-3页
[19]杨德仁,薛梅,顾君忠.Web Services核心协议与实施研究.计算机应用系统.2005,(1):33-36页
[20]王楠,刘心雄,陈和平.Web Service技术研究.计算机与数字工程.2006,34(7):88-90页
[21] Kunz, T..The influence of different workload descriptions on a heuristic load balancing scheme[J].IEEE Transactions on Software Engineering, 1991, 17(7): 725-730P
[22] Kameda H, Fathy S, Ryu I et a1.A performance comparison of dynamic vs. static load balancing policies in a mainframe-personal computer network model.Proceedings of IEEE CDC2000, Sydney, 2000: 1415-1420P
[23]陈涛,陈启买.分布式计算机系统负载平衡研究.计算机技术与发展.2006,16(5):33-35页
[24] Eager, D.L., E.D.Lazowska, and J.Zahorijan . A comparison of receiver-initiated and sender-initiated adaptive load sharing[J].performance Evaluation, 1986, 1(3): 53-68P
[25] Lin, F.C.H. and R.M.Keller . The gradient model load balancing method.IEEE Transactions on Software Engineering, 1987, 13(1): 32-28P
[26] Bryant, R.M. and R.A.Finkel . A stable distributed scheduling algorithm.Proc. of the 2~(nd) int’1 Conf. on Distributed Computing Systems, 1981: 314-323P
[27]李冬梅,施海虎.负载平衡调度问题的一般模型研究.计算机工程与应用.2007,43(8):121-125页
[28] Singhal, M. and N.G. Shivaratri.Advanced Concepts in Operating Systems: Distributed,Database, S. and Multiprocessors Operating Systems, McGraw-Hill, Inc., 1994
[29] Leff, A. and P.S. Yu.An adaptive strategy for load sharing in distributed database environments with information lags.Journal of Parallel and Distributed Computing, 1991, 13(1): 91-103P
[30]彭玲玲.改进的分布式动态自适应负载平衡算法探讨.电脑学习.2009,(3):134-135页
[31]罗拥军,李晓乐,孙如祥.负载均衡算法综述.科技情报开发与经济.2008,18(23):134-135页
[32]王友良,叶柏龙.分布式系统中动态负载平衡的研究.科学技术与工程.2005,5(9):572-574页
[33]李岱松,刘勇,潘成胜.一种负载均衡算法的研究与实现.沈阳理工大学学报.2005,24(3):47-48页
[34]张宇晴,佟振声,胡旦华.分布式系统中动态负载平衡算法的研究.计算机仿真.2003,20(9):69-70页
[35] Suen, T. T. Y. and J. S. K. Wong.Efficient task migration algorithm for distributed systems.IEEE Transactions on Parallel and Distributed Systems, 1992, 3(4): 488-499P
[36]刘建,李绪志.一种动态负载均衡机制的研究与实现.计算机工程与应用.2006,(2):142-144页