基于Agent技术的网格数据复制管理研究
|
摘要
网格把整个互联网整合成一台巨大的超级计算机，实现各种资源的全面共享。网格中的数据复制管理可以减轻单个服务结点的通信和访问压力，对于减轻频繁的远程访问对网络带宽资源的消耗、平衡负载等方面也具有重要的作用。
Agent是具有自治性、社会性、主动性和适应性的智能实体。基于Agent的方法将使得网格的管理更加智能和方便。
目前，将Agent技术应用于Globus网格的数据复制管理上的并不多见，很多问题有待解决。
本文探讨了在网格环境下，如何利用Agent技术，进行有效的数据复制管理。本文运用Agent的智能性等特点，针对数据复制管理的几个方面提出了改进思想：动态管理权方式进行数据主副本一致性的维护，扩展了存储与串联复制结合的复制策略，在目录设置上引入了全局复制目录。应用这些思想，设计了一个网格环境下进行数据复制管理的多Agent管理体系，并通过对Agent的具体实现以及Globus Toolkit 3.0的相关基础服务的调用的探讨，说明了该体系在实现上的可行性以及其具体的实现方法。最后，在假定的网格的网络环境和数据访问状况下，模拟分析了该系统进行复制管理的过程，并在结果数据基础上，对系统以及其中的改进设计进行了简单评测。
Grid computing, which integrates worldwide Internet resource into virtual supercomputer, accomplish every kind of resources to full share in global. Data replication management in the grid can reduce the communication cost on solitary service site. It also has good effect on reducing the bandwidth consumption caused by frequent remote access and balancing the load of global grid.
Agent is a kind of intelligent entity and is autonomous, social, active and adaptable. Agent technique can provide an ideal approach to solve question in the grid computing. At present, agent technique is seldom well used in data replication management on Grid associated with Globus technology.Many problems need to be solved.
This paper discusses how to apply agent technique on data replication management on Grid associated with Globus technology. It proposes several ideas on keeping the Consistency between original file and it's copies, application of replication strategies and catalog setting. And based on these ideas, it designs a data replication management system on Grid by the use of agent technique. After discussing the implementation of Agent and the calling of Base Services in Globus Toolkit 3.0,the paper show the feasibility of the system and how to bring it to pass. At the end of the paper, we simulate the procedures of data replication managed by the designed system, in an assumed Grid environment. And the replication management system designed was analyzed based on these simulate procedures.
Agent是具有自治性、社会性、主动性和适应性的智能实体。基于Agent的方法将使得网格的管理更加智能和方便。
目前，将Agent技术应用于Globus网格的数据复制管理上的并不多见，很多问题有待解决。
本文探讨了在网格环境下，如何利用Agent技术，进行有效的数据复制管理。本文运用Agent的智能性等特点，针对数据复制管理的几个方面提出了改进思想：动态管理权方式进行数据主副本一致性的维护，扩展了存储与串联复制结合的复制策略，在目录设置上引入了全局复制目录。应用这些思想，设计了一个网格环境下进行数据复制管理的多Agent管理体系，并通过对Agent的具体实现以及Globus Toolkit 3.0的相关基础服务的调用的探讨，说明了该体系在实现上的可行性以及其具体的实现方法。最后，在假定的网格的网络环境和数据访问状况下，模拟分析了该系统进行复制管理的过程，并在结果数据基础上，对系统以及其中的改进设计进行了简单评测。
Grid computing, which integrates worldwide Internet resource into virtual supercomputer, accomplish every kind of resources to full share in global. Data replication management in the grid can reduce the communication cost on solitary service site. It also has good effect on reducing the bandwidth consumption caused by frequent remote access and balancing the load of global grid.
Agent is a kind of intelligent entity and is autonomous, social, active and adaptable. Agent technique can provide an ideal approach to solve question in the grid computing. At present, agent technique is seldom well used in data replication management on Grid associated with Globus technology.Many problems need to be solved.
This paper discusses how to apply agent technique on data replication management on Grid associated with Globus technology. It proposes several ideas on keeping the Consistency between original file and it's copies, application of replication strategies and catalog setting. And based on these ideas, it designs a data replication management system on Grid by the use of agent technique. After discussing the implementation of Agent and the calling of Base Services in Globus Toolkit 3.0,the paper show the feasibility of the system and how to bring it to pass. At the end of the paper, we simulate the procedures of data replication managed by the designed system, in an assumed Grid environment. And the replication management system designed was analyzed based on these simulate procedures.
引文
[1]都志辉 陈渝 刘鹏．网格计算．北京：清华大学出版社，2002．10
[2]肖侬．数据网格研究现状．Http://www.gridhome.com
[3]刘鹏．我国网格研究现状．http://www.gridhome.com
[4]许文韬．网络技术综述．微型电脑应用．2002年第18卷第6期
[5]李伟 谈恩华．Globus项目进展和技术水平分析．
[6]张艳莉，徐英卓，李洪涛．基于多Agent的远程教学系统模型的研究．微机发展 No．10 Oct．2003 Vol．13
[7]陈萍，余华山，王彬，许卓群．网格计算环境Globus介绍．计算机应用研究．2003年文章编号：100123695(2003)0820096203
[8]赵瑞清，王晖，邱涤虹．知识表示与推理．1991
[9]庞丽萍 陈勇．数据网格环境下的一种动态备份模型．华中科技大学计算机科学与技术学院
[10]肖侬．编织“数据网格”——实现数据网格的关键技术．计算机世界报．第40期B9、B10
[11]朱子玉．OGSA安装与程序设计．清华大学计算机系高性能所网格研究组
[12]贾永洁 金海 易川江 韩宗芬 吴松．存储虚拟化系统的元数据副本一致性管理模型．中图分类号：TP393．稿件编号：23254
[13]庞丽萍 陈勇．网格环境下数据副本创建策略．华中科技大学计算机科学与技术学院
[14]李彬．用GT3开发网格服务．IBM中国网站．2003年9月
[15]林新华 赵靖 翁楚良 孙犀利．GT3安装完全指南．上海交通大学—IBM网格中心
[16]陈冠岭、繁晓聪、徐殿祥、郑国梁．Agent的通讯语言—KQML．计算机科学 1998 Vol．25．No．26
[17]武秀川，胡亮，鞠九滨．数据网格的数据管理策略．小型微型计算机系统．2004(1)第25卷第1期
[18]W. H. Bell, D. G. Cameron, L. Capozza, A. P. Millar, K. Stockinger, F. Zini. " Studying Dynamic Grid Optimization Algorithms for File Replication"
[19]Houda Lamehamedi, Boleslaw Szymanski, and Zujun Shentu. Data Replication Strategies in Grid Environments.
[20]Keith Rang Nathan and Ian Foster. Identifying Dynamic Replication Strategies for a High-Performance Data Grid.
[21]w. Allcockl, A.Chervenak, I. Poster, L.Pearlman, V.Welch, M. Wide. Globus Toolkit Support of Distributed Data—Intensive Science. Beijing: Computing in High Energy and Nuclear
Physics, 2001
[22]Junwei Cao, Darren J. Kerbyson, and Graham R. Nudd. Performance Evaluation of an Agent-Based Resource Management Infrastructure for Grid Computing.
[23]William Allcock, Ian Foster, and Steven Tuecke. Protocols and Services for Distributed Data-Intensive Science
[24]Joshy Joseph. Giobus ToolKit 3.0 and OGSI architecture-overview.http://www-106.ibm.com/developerworks/grid/library/gr-gt3/index.html
[25]Borja Sotomayor. The Globus Toolkit 3 Programmer's Tutorial. http://www. Globustoolkit.org. 28-Jul-2003
[26]Ian Foster. Internet Computing and the Emerging Grid. Nature Web Matters. 7 December 2000
[27]Ian Foster, Carl Kesselman. The Globus Project: A Status Report
[28]Wooldridge M J. Agent-Based Software Engineering. IEEE Transaction on Software Engineering, 1997, 144, (1)
[29]William H. Bell, David G. Cameron, Ruben Carvajal-Schiaffino, A. Paul Millar, Kurt Stockinger, Floriano Zini. Evaluation of an Economy-Based File Replication Strategy for a Data Grid
[30]James Frey, Todd Tannenbaum, Miron Livny Ian Foster, Steven Tuecke. Condor-G: A Computation Management Agent for Multi-Institutional Grids
[31]Heinz Stockinger, Asad Samar, Bill Allcock, Ian Foster, Koen, Brian Tierney. Holtman File and Object Replication in Data Grids.
[32]Michael Wooldridge, Nicholas R. Jennings, Agent Theories, Architectures, and Languages: a Survey, in Wooldridge and Jennings Eds., Intelligent Agents, Berlin: Springer-Verlag, 1-22, p2, 1995
[33]James Frey, Todd Tannenbaum, Miron Livny Ian Foster, Steven Tuecke. Condor-G: A Computation Management Agent for Multi-Institutional Grids
[34]F.Manola, C.Thompson, Characterizing the agent grid. http://www.objs.com/agility/techreports/990623-characterizing-the-agentgrid. html, June 1999.
[35]B.J. Overeinder, N.J.E. Wijngaards, M. van Steen, and F.M.T. Brazier Multi-Agent Support for Internet-Scale Grid Management 2002
[36]Ian Foster, Carl Kesselman. Globus:A Metacomputing Infrastructure Toolkit. http://www.globus.org/
[37]Ron Oldfield. Summary of Existing and Developing Data Grids. March 2002
[2]肖侬．数据网格研究现状．Http://www.gridhome.com
[3]刘鹏．我国网格研究现状．http://www.gridhome.com
[4]许文韬．网络技术综述．微型电脑应用．2002年第18卷第6期
[5]李伟 谈恩华．Globus项目进展和技术水平分析．
[6]张艳莉，徐英卓，李洪涛．基于多Agent的远程教学系统模型的研究．微机发展 No．10 Oct．2003 Vol．13
[7]陈萍，余华山，王彬，许卓群．网格计算环境Globus介绍．计算机应用研究．2003年文章编号：100123695(2003)0820096203
[8]赵瑞清，王晖，邱涤虹．知识表示与推理．1991
[9]庞丽萍 陈勇．数据网格环境下的一种动态备份模型．华中科技大学计算机科学与技术学院
[10]肖侬．编织“数据网格”——实现数据网格的关键技术．计算机世界报．第40期B9、B10
[11]朱子玉．OGSA安装与程序设计．清华大学计算机系高性能所网格研究组
[12]贾永洁 金海 易川江 韩宗芬 吴松．存储虚拟化系统的元数据副本一致性管理模型．中图分类号：TP393．稿件编号：23254
[13]庞丽萍 陈勇．网格环境下数据副本创建策略．华中科技大学计算机科学与技术学院
[14]李彬．用GT3开发网格服务．IBM中国网站．2003年9月
[15]林新华 赵靖 翁楚良 孙犀利．GT3安装完全指南．上海交通大学—IBM网格中心
[16]陈冠岭、繁晓聪、徐殿祥、郑国梁．Agent的通讯语言—KQML．计算机科学 1998 Vol．25．No．26
[17]武秀川，胡亮，鞠九滨．数据网格的数据管理策略．小型微型计算机系统．2004(1)第25卷第1期
[18]W. H. Bell, D. G. Cameron, L. Capozza, A. P. Millar, K. Stockinger, F. Zini. " Studying Dynamic Grid Optimization Algorithms for File Replication"
[19]Houda Lamehamedi, Boleslaw Szymanski, and Zujun Shentu. Data Replication Strategies in Grid Environments.
[20]Keith Rang Nathan and Ian Foster. Identifying Dynamic Replication Strategies for a High-Performance Data Grid.
[21]w. Allcockl, A.Chervenak, I. Poster, L.Pearlman, V.Welch, M. Wide. Globus Toolkit Support of Distributed Data—Intensive Science. Beijing: Computing in High Energy and Nuclear
Physics, 2001
[22]Junwei Cao, Darren J. Kerbyson, and Graham R. Nudd. Performance Evaluation of an Agent-Based Resource Management Infrastructure for Grid Computing.
[23]William Allcock, Ian Foster, and Steven Tuecke. Protocols and Services for Distributed Data-Intensive Science
[24]Joshy Joseph. Giobus ToolKit 3.0 and OGSI architecture-overview.http://www-106.ibm.com/developerworks/grid/library/gr-gt3/index.html
[25]Borja Sotomayor. The Globus Toolkit 3 Programmer's Tutorial. http://www. Globustoolkit.org. 28-Jul-2003
[26]Ian Foster. Internet Computing and the Emerging Grid. Nature Web Matters. 7 December 2000
[27]Ian Foster, Carl Kesselman. The Globus Project: A Status Report
[28]Wooldridge M J. Agent-Based Software Engineering. IEEE Transaction on Software Engineering, 1997, 144, (1)
[29]William H. Bell, David G. Cameron, Ruben Carvajal-Schiaffino, A. Paul Millar, Kurt Stockinger, Floriano Zini. Evaluation of an Economy-Based File Replication Strategy for a Data Grid
[30]James Frey, Todd Tannenbaum, Miron Livny Ian Foster, Steven Tuecke. Condor-G: A Computation Management Agent for Multi-Institutional Grids
[31]Heinz Stockinger, Asad Samar, Bill Allcock, Ian Foster, Koen, Brian Tierney. Holtman File and Object Replication in Data Grids.
[32]Michael Wooldridge, Nicholas R. Jennings, Agent Theories, Architectures, and Languages: a Survey, in Wooldridge and Jennings Eds., Intelligent Agents, Berlin: Springer-Verlag, 1-22, p2, 1995
[33]James Frey, Todd Tannenbaum, Miron Livny Ian Foster, Steven Tuecke. Condor-G: A Computation Management Agent for Multi-Institutional Grids
[34]F.Manola, C.Thompson, Characterizing the agent grid. http://www.objs.com/agility/techreports/990623-characterizing-the-agentgrid. html, June 1999.
[35]B.J. Overeinder, N.J.E. Wijngaards, M. van Steen, and F.M.T. Brazier Multi-Agent Support for Internet-Scale Grid Management 2002
[36]Ian Foster, Carl Kesselman. Globus:A Metacomputing Infrastructure Toolkit. http://www.globus.org/
[37]Ron Oldfield. Summary of Existing and Developing Data Grids. March 2002