一种新的基于地层调度协议的网格模型及其资源预留技术
|
摘要
作为网格系统最重要构成部分之一的网络资源管理是对发展网格系统极具意义的技术。网格资源管理的内容范围广泛,包括了网络地点资源、网络组织资源、网络资源分配、网络资源调度等多方面。众多单一的网络资源能够通过网格管理系统集成一个整体,统一进行处理过度。先进的网络资源管理软件能够合理的配置资源、运用资源,对实现网络资源共享有着重要意义。
本文以传统网格传统调度模型P2P-Grid为对象进行分析,针对现存的几个问题,提出构建新的调度模型Group-Grid。
Group-Grid作为新的网格结构模型,能够根据网络资源提供者的不同将网格资源分为不同的小区,通过对各小区信息资源的快速搜索得到资源共享的目的。Group-Grid相比P2P-Grid,信息资源搜索速度更快,搜索过程带来的网络负担较小,并且容易调整,扩展性强。Group-Grid的资源预留策略不仅具有普遍性的预留策略,还具备了任务预留及其级别判定等多项功能,这些功能提高了模型资源预留成功度,有利于保留客户资源的QOS,最终实现模型在资源预留上的灵敏性和可靠性。除此之前,本文针对传统调度模型P2P-Grid无法平衡利用网络资源的缺陷,提出了改进式的网络资源预留计算法,即通过对即时任务快速的转化为预留任务,从而减少网络即时任务等待处理的时间,减轻网络任务处理的负担,最终实现网络资源共享高速度。本文通过相关模型的仿真数据说明,资源预留策略与即时任务处理速度相比提高了百分之十五的成功率。
Resource management of grid system is one of the most important technologies, and it is of great theoretical practical value for the development of grid computing, location, organization, scheduling, allocation, confirming the discovery process, creating and preparing the necessary resources.It provides advanced management features and concepts, the clusters can be used as a single resource, administrator according to predetermined criteria, through resource management software, ensure the reasonable configuration of resource and use, achieves resources sharing.
The main work of this paper is based on the traditional hierarchical scheduling model (P2P-Grid), proposes a new grid structure model (Group-Grid), to solve the traditional grid structure model in the surviving several main problems.
This paper presents the model cf Group-Grid, is the grid resources according to the resource provider is divided into different compartments.Model of service discovery module to realize the purpose of quick search resources.Improved P2P-Grid using pan Hong Fa led to the network load increases, the disadvantage of the long search cycle.But also has strong expansibility, easy adjustment etc.Based on the structure model of resource reservation strategy, is available in the general reserve strategy based on task, put forward block, reservation cancel and reserved level concepts, so as to ensure the reliability and flexibility of the reservation of resources, and improve the success rate of Resource Reservation Reservation and customer QoS.In addition, the traditional resource reservation algorithm can balance the network overall resource utilization is insufficient, the existing balance tasks of resource reservation algorithm, this paper proposes an improved resource reservation algorithm, the immediate task is transformed into the reservation task submitted to the entire network to reduce the immediate task of the average waiting time, and can realization of whole network load balance, improve the utilization rate of resources in grid.The simulation results show that, the resource reservation strategy than the old strategy in the success rate improved by15%, in the cyber source utilization rate is improved by5%.
本文以传统网格传统调度模型P2P-Grid为对象进行分析,针对现存的几个问题,提出构建新的调度模型Group-Grid。
Group-Grid作为新的网格结构模型,能够根据网络资源提供者的不同将网格资源分为不同的小区,通过对各小区信息资源的快速搜索得到资源共享的目的。Group-Grid相比P2P-Grid,信息资源搜索速度更快,搜索过程带来的网络负担较小,并且容易调整,扩展性强。Group-Grid的资源预留策略不仅具有普遍性的预留策略,还具备了任务预留及其级别判定等多项功能,这些功能提高了模型资源预留成功度,有利于保留客户资源的QOS,最终实现模型在资源预留上的灵敏性和可靠性。除此之前,本文针对传统调度模型P2P-Grid无法平衡利用网络资源的缺陷,提出了改进式的网络资源预留计算法,即通过对即时任务快速的转化为预留任务,从而减少网络即时任务等待处理的时间,减轻网络任务处理的负担,最终实现网络资源共享高速度。本文通过相关模型的仿真数据说明,资源预留策略与即时任务处理速度相比提高了百分之十五的成功率。
Resource management of grid system is one of the most important technologies, and it is of great theoretical practical value for the development of grid computing, location, organization, scheduling, allocation, confirming the discovery process, creating and preparing the necessary resources.It provides advanced management features and concepts, the clusters can be used as a single resource, administrator according to predetermined criteria, through resource management software, ensure the reasonable configuration of resource and use, achieves resources sharing.
The main work of this paper is based on the traditional hierarchical scheduling model (P2P-Grid), proposes a new grid structure model (Group-Grid), to solve the traditional grid structure model in the surviving several main problems.
This paper presents the model cf Group-Grid, is the grid resources according to the resource provider is divided into different compartments.Model of service discovery module to realize the purpose of quick search resources.Improved P2P-Grid using pan Hong Fa led to the network load increases, the disadvantage of the long search cycle.But also has strong expansibility, easy adjustment etc.Based on the structure model of resource reservation strategy, is available in the general reserve strategy based on task, put forward block, reservation cancel and reserved level concepts, so as to ensure the reliability and flexibility of the reservation of resources, and improve the success rate of Resource Reservation Reservation and customer QoS.In addition, the traditional resource reservation algorithm can balance the network overall resource utilization is insufficient, the existing balance tasks of resource reservation algorithm, this paper proposes an improved resource reservation algorithm, the immediate task is transformed into the reservation task submitted to the entire network to reduce the immediate task of the average waiting time, and can realization of whole network load balance, improve the utilization rate of resources in grid.The simulation results show that, the resource reservation strategy than the old strategy in the success rate improved by15%, in the cyber source utilization rate is improved by5%.
引文
[1]Catlett C, Smarr L. Metacomputing [J]. Communications of the ACM,1992, 35(6):44-52
[2]Foster I, Kesselman C, Nick J M, et al. The physiology of the grid:An open grid services architecture for distributed systems integration [J]. Grid computing:, 2002,32(8):217-250
[3]Foster I, Kesselman C, Tuecke S. The anatomy of the grid:Enabling scalable virtual organizations [J]. International Journal of High-performance Computing Application,2001,15(3):200-222
[4]Foster I, Kesselman C. The globus project:A status report [J]. Future Generation Computer Systems,1999,15(5-6):607-621
[5]Czajkowski K, Foster I, Karonis N, et al. A resource management architecture for metacomputing systems [J]. Lecture Notes in Computer Science 1459. 1998,13(2).62-82
[6]李云,隆克平,赵为粮,陈前斌,吴诗其D-PRMA：一种新的无线Ad Hoc网络资源预留MAC机制.电子与信息学报,2005,27(06)：967-972.
[7]Min R, Maheswaran M. Scheduling co-reservations with priorities in grid computing systems[J]. IEEE Computer Society Press,2002,16(5):266-269.
[8]Mateescu G Quality of service on the grid via metascheduling with resource co-scheduling and co-reservation [J]. International Journal of High Performance Computing Applications,2003,17(3):209-218
[9]Foster I, Fidler M, Roy A, et al. End-to-end quality of service for high-end applications [J]. Computer Communications,2004,27(14):1375-1388
[10]Czajkowski K, Foster I, Karonis N, et al. A resource management architecture for metacomputing systems[A]. Feitelson D G, Rudolph L. Lecture Notes in Computer Science 1459[C]. Berlin:Springer-Verlag,1998:62-82.
[11]Foster I, Kesselman C. Globus:A Metacomputing Infrastructure Toolkit [J]. Intl J. Supercompute Applications.1997,11(2):115-128
[12]Grimshaw A. S, Wulf W. A. The Legion Vision of a World wide Virtual Computer [J]. Communications o the ACM,1997,40(1):39-45
[13]Litzkow M. J, Livny M, Mutka M. W. Condor-A hunter of idle workstations [C]. 8th International Conference on Distributed Computing Systems, 1988,32(7):104-111
[14]Segal B. Grid Computing:The European Data Project [C], IEEE Nuclear Science Symposium and Medica Imaging Conference,2000,3(2):15-20
[15]毛晓峰.无线传感器网络与车载自组网路由技术研究[D].上海：复旦大学, 2008：34.
[16]Sato M, Nakada H, Sekiguchi S, et al. A Network based Information Library for a Globa World-Wide Computing Infrastructure [J]. Proceedings of HPCN,1997:491-502
[17]徐志伟,李伟.织女星网格的体系结构研究[J].计算机研究与发展,2002,39(8)：32-34
[18]Jon MacLaren. Advance reservations:state of the art [J]. the GGF GRAAP-WG, 2002,10(4):24-28
[19]李波.支持网格资源预留的作业调度算法研究[D].武汉：华中科技大学,2005：26
[20]P Shenoy, H Vin.Cello. A disk scheduling framework for next generation operating systems [J].ACM SIGMERICS'98,1998,5(12):44-55
[21]L Zhang, S Deering, D Estrin, Shenker, D Zappala. Resource Reservation Protocol(RSVP) [J].IEEE Network,1993,25(5):253-257.
[22]H Chu, K Nahrstedt. CPU service classes for multimedia applications [J]. On Multimedia Computing and System,1999,35(3):296-301.
[23]G Garimella. Advance CPU Reservations with the Dynamic Soft Real-Time Scheduler [D]. USA:University of Illinois at Urbana-Champaign,1999:52.
[24]K Kim, K Nahrstedt. A Resource Broker Model with Integrated Reservation Scheme[J]. IEEE International Conference on Multimedia and Expo2000, 2000,32(12):859-862.
[25]叶从欢P2P-Grid模型及其环境下资源管理关键技术研究[D].成都：电子科技大学,2004：20
[26]裘楷.一种面向服务的P2P网格资源管理模型[D].西安：西安电子科技大学,2007：13
[27]Barry G, Lawson, Evgenia Smirni. Multiple-queue Backfill Scheduling with Priorities and Reservation for Parallel Systems[J].8th Workshop on Job Scheduling Strategies for Parallel processing,2002,5(8),45-48.
[28]F Azzedin and M Maheswaran. Synchronous Queuing:A Co-allocation Mechanism for Multimedia Enabled Grids[J]. Thirteenth IASTED International Conference on Parallel and Distributed Computing System,2001,3(4):27-32.
[29]Gary Hoo, W Johnston. QoS as middleware:bandwidth reservation system design. The Eighth IEEE International Symposium on High Performance Distributed Computing,1999,1(1):355-356.
[30]D Xu, K Nahrstedt, A Viswanathan, D Wichadkul. QoS and contention-aware multi-resource reservation. Proceedings of the 9th IEEE International Symposium on High Performance Distributed Computing,2001,34(6):3-10.
[31]蒲静.基于任务可分的网格资源预留机制[J].计算机工程与应用,2008,44(12)：118-120.
[32]李冉,于炯,侯勇.在共享环境下基于时间预测的资源预留[J].现代计算机,2008,29(1)：4-6.
[33]胡志刚,谌任.一种兼顾即时任务的资源预留机制[J].计算机工程,2006,32(12)：60-61.
[34]丁长松,胡志刚,胡周君.一种面向效益均衡的网格资源预留机制[J].小型微型计算机系统,2010,31(3)：413-417.
[35]黄智兴.网格环境下基于经济机制的资源预留方法研究[D],重庆：西南大学,2006：25.
[36]田东,陈蜀宇,吴长泽.基于间隔型双边拍卖的网格资源预留算法[J].计算机应用,2007,27(1)：15-18.
[37]丁长松,胡志刚,胡周君.一种面向随机性负载的计算网格资源预留定价策略[J].计算机应用,2009,29(6)：1711-1714.
[38]李静,韩起云.远程教育网格资源预留模型[J].计算机工程,2008,34(20)：275-276.
[39]Foster I, Kesselman C. The Grid Blueprint for a Future Computing Infrastructure[D].San Mateo:Morgan Kaufmann Publishers,1999:325.
[40]Foster I, Kesselman C, Lee C. A Distributed Resource Management Architecture That Supports Advance Reservations and Co-allocation[J]. International Workshop on Quality of Service,1999,24(7):27-36.
[41]Dana Moore, John Hebeler,苏忠,战晓雷PEER-TO-PEER[M]北京：清华大学出版社,2003：165.
[42]Wolfgang Hoschek. A Unified Peer-to-Peer Database Framework for Scalable Service and ResourceDiscovery. CERN,2002,34(2):42-46.
[43]Ian Foster, Adriana Ianmitch. A Peer-to-Peer Approach to Resource Location in Grid Environments. High Performance Distributed Computing,2002, 6(11):419-425.
[44]Shrideep Pallickara, Geoffrey Fox, Narada Brokering.A Distributed Middleware Framework and Architecture for Enabling Durable Peer-to-Peer Grids[J]. Proceedings of ACM/IFIP/USENIX International Middleware Conference,2003, 2(15):146-152.
[45]曹亮,王茜,卢菁.XML数据在关系数据库中存储和检索的研究和实现[J].东南大学学报(自然科学版),2002,31(01)：124-127.
[46]Smith V, Foster I, Taylor V. Scheduling with Advanced Reservation[J].International Parallel and Distributed Processing Symposium,2000, 5(18):127-132.
[47]Sulistio A, Buyya R.A Grid Simulation Infrastructure Supporting Advance Reservation[A]. Proceedings of the 16th International Conference on Parallel and Distributed Computing and Systems[C]. USA:MIT Cambridge,2004:1-7.
[2]Foster I, Kesselman C, Nick J M, et al. The physiology of the grid:An open grid services architecture for distributed systems integration [J]. Grid computing:, 2002,32(8):217-250
[3]Foster I, Kesselman C, Tuecke S. The anatomy of the grid:Enabling scalable virtual organizations [J]. International Journal of High-performance Computing Application,2001,15(3):200-222
[4]Foster I, Kesselman C. The globus project:A status report [J]. Future Generation Computer Systems,1999,15(5-6):607-621
[5]Czajkowski K, Foster I, Karonis N, et al. A resource management architecture for metacomputing systems [J]. Lecture Notes in Computer Science 1459. 1998,13(2).62-82
[6]李云,隆克平,赵为粮,陈前斌,吴诗其D-PRMA：一种新的无线Ad Hoc网络资源预留MAC机制.电子与信息学报,2005,27(06)：967-972.
[7]Min R, Maheswaran M. Scheduling co-reservations with priorities in grid computing systems[J]. IEEE Computer Society Press,2002,16(5):266-269.
[8]Mateescu G Quality of service on the grid via metascheduling with resource co-scheduling and co-reservation [J]. International Journal of High Performance Computing Applications,2003,17(3):209-218
[9]Foster I, Fidler M, Roy A, et al. End-to-end quality of service for high-end applications [J]. Computer Communications,2004,27(14):1375-1388
[10]Czajkowski K, Foster I, Karonis N, et al. A resource management architecture for metacomputing systems[A]. Feitelson D G, Rudolph L. Lecture Notes in Computer Science 1459[C]. Berlin:Springer-Verlag,1998:62-82.
[11]Foster I, Kesselman C. Globus:A Metacomputing Infrastructure Toolkit [J]. Intl J. Supercompute Applications.1997,11(2):115-128
[12]Grimshaw A. S, Wulf W. A. The Legion Vision of a World wide Virtual Computer [J]. Communications o the ACM,1997,40(1):39-45
[13]Litzkow M. J, Livny M, Mutka M. W. Condor-A hunter of idle workstations [C]. 8th International Conference on Distributed Computing Systems, 1988,32(7):104-111
[14]Segal B. Grid Computing:The European Data Project [C], IEEE Nuclear Science Symposium and Medica Imaging Conference,2000,3(2):15-20
[15]毛晓峰.无线传感器网络与车载自组网路由技术研究[D].上海：复旦大学, 2008：34.
[16]Sato M, Nakada H, Sekiguchi S, et al. A Network based Information Library for a Globa World-Wide Computing Infrastructure [J]. Proceedings of HPCN,1997:491-502
[17]徐志伟,李伟.织女星网格的体系结构研究[J].计算机研究与发展,2002,39(8)：32-34
[18]Jon MacLaren. Advance reservations:state of the art [J]. the GGF GRAAP-WG, 2002,10(4):24-28
[19]李波.支持网格资源预留的作业调度算法研究[D].武汉：华中科技大学,2005：26
[20]P Shenoy, H Vin.Cello. A disk scheduling framework for next generation operating systems [J].ACM SIGMERICS'98,1998,5(12):44-55
[21]L Zhang, S Deering, D Estrin, Shenker, D Zappala. Resource Reservation Protocol(RSVP) [J].IEEE Network,1993,25(5):253-257.
[22]H Chu, K Nahrstedt. CPU service classes for multimedia applications [J]. On Multimedia Computing and System,1999,35(3):296-301.
[23]G Garimella. Advance CPU Reservations with the Dynamic Soft Real-Time Scheduler [D]. USA:University of Illinois at Urbana-Champaign,1999:52.
[24]K Kim, K Nahrstedt. A Resource Broker Model with Integrated Reservation Scheme[J]. IEEE International Conference on Multimedia and Expo2000, 2000,32(12):859-862.
[25]叶从欢P2P-Grid模型及其环境下资源管理关键技术研究[D].成都：电子科技大学,2004：20
[26]裘楷.一种面向服务的P2P网格资源管理模型[D].西安：西安电子科技大学,2007：13
[27]Barry G, Lawson, Evgenia Smirni. Multiple-queue Backfill Scheduling with Priorities and Reservation for Parallel Systems[J].8th Workshop on Job Scheduling Strategies for Parallel processing,2002,5(8),45-48.
[28]F Azzedin and M Maheswaran. Synchronous Queuing:A Co-allocation Mechanism for Multimedia Enabled Grids[J]. Thirteenth IASTED International Conference on Parallel and Distributed Computing System,2001,3(4):27-32.
[29]Gary Hoo, W Johnston. QoS as middleware:bandwidth reservation system design. The Eighth IEEE International Symposium on High Performance Distributed Computing,1999,1(1):355-356.
[30]D Xu, K Nahrstedt, A Viswanathan, D Wichadkul. QoS and contention-aware multi-resource reservation. Proceedings of the 9th IEEE International Symposium on High Performance Distributed Computing,2001,34(6):3-10.
[31]蒲静.基于任务可分的网格资源预留机制[J].计算机工程与应用,2008,44(12)：118-120.
[32]李冉,于炯,侯勇.在共享环境下基于时间预测的资源预留[J].现代计算机,2008,29(1)：4-6.
[33]胡志刚,谌任.一种兼顾即时任务的资源预留机制[J].计算机工程,2006,32(12)：60-61.
[34]丁长松,胡志刚,胡周君.一种面向效益均衡的网格资源预留机制[J].小型微型计算机系统,2010,31(3)：413-417.
[35]黄智兴.网格环境下基于经济机制的资源预留方法研究[D],重庆：西南大学,2006：25.
[36]田东,陈蜀宇,吴长泽.基于间隔型双边拍卖的网格资源预留算法[J].计算机应用,2007,27(1)：15-18.
[37]丁长松,胡志刚,胡周君.一种面向随机性负载的计算网格资源预留定价策略[J].计算机应用,2009,29(6)：1711-1714.
[38]李静,韩起云.远程教育网格资源预留模型[J].计算机工程,2008,34(20)：275-276.
[39]Foster I, Kesselman C. The Grid Blueprint for a Future Computing Infrastructure[D].San Mateo:Morgan Kaufmann Publishers,1999:325.
[40]Foster I, Kesselman C, Lee C. A Distributed Resource Management Architecture That Supports Advance Reservations and Co-allocation[J]. International Workshop on Quality of Service,1999,24(7):27-36.
[41]Dana Moore, John Hebeler,苏忠,战晓雷PEER-TO-PEER[M]北京：清华大学出版社,2003：165.
[42]Wolfgang Hoschek. A Unified Peer-to-Peer Database Framework for Scalable Service and ResourceDiscovery. CERN,2002,34(2):42-46.
[43]Ian Foster, Adriana Ianmitch. A Peer-to-Peer Approach to Resource Location in Grid Environments. High Performance Distributed Computing,2002, 6(11):419-425.
[44]Shrideep Pallickara, Geoffrey Fox, Narada Brokering.A Distributed Middleware Framework and Architecture for Enabling Durable Peer-to-Peer Grids[J]. Proceedings of ACM/IFIP/USENIX International Middleware Conference,2003, 2(15):146-152.
[45]曹亮,王茜,卢菁.XML数据在关系数据库中存储和检索的研究和实现[J].东南大学学报(自然科学版),2002,31(01)：124-127.
[46]Smith V, Foster I, Taylor V. Scheduling with Advanced Reservation[J].International Parallel and Distributed Processing Symposium,2000, 5(18):127-132.
[47]Sulistio A, Buyya R.A Grid Simulation Infrastructure Supporting Advance Reservation[A]. Proceedings of the 16th International Conference on Parallel and Distributed Computing and Systems[C]. USA:MIT Cambridge,2004:1-7.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |