Administration of grid resources is a time consuming and often tedious job. Most administrative requests are predictable, and in general, handling them requires knowledge of the local resources and the requester. In this paper we discuss a system to provide automated support for administrative requests, such as resource reservation and user account management. We propose using trust metrics to help judge the merits and suitability of each request. We outline how these metrics can be implemented using trust management techniques into a practical system we call GridAdmin.
{"title":"GridAdmin: decentralising grid administration using trust management","authors":"T. Quillinan, B. Clayton, S. Foley","doi":"10.1109/ISPDC.2004.28","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.28","url":null,"abstract":"Administration of grid resources is a time consuming and often tedious job. Most administrative requests are predictable, and in general, handling them requires knowledge of the local resources and the requester. In this paper we discuss a system to provide automated support for administrative requests, such as resource reservation and user account management. We propose using trust metrics to help judge the merits and suitability of each request. We outline how these metrics can be implemented using trust management techniques into a practical system we call GridAdmin.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73639459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
One of the barriers that prevents the expansion and adoption of grid technologies is the lack of a standard programming paradigm to port existing applications among different environments. The distributed resource management application API (DRMAA) has been proposed to aid the rapid development and distribution of these applications across the grid. In this paper we present the first implementation of the DRMAA standard on a Globus-based testbed, and show its suitability to express typical scientific applications. As a case of study, we will consider in this paper the implementation of the NAS grid benchmarks with DRMAA. The DRMAA routines are supported by the functionality offered by the GridWay framework, that provides the runtime mechanisms needed for transparently executing jobs on a dynamic grid environment.
{"title":"Execution of typical scientific applications on Globus-based grids","authors":"José Herrera, E. Huedo, R. Montero, I. Llorente","doi":"10.1109/ISPDC.2004.23","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.23","url":null,"abstract":"One of the barriers that prevents the expansion and adoption of grid technologies is the lack of a standard programming paradigm to port existing applications among different environments. The distributed resource management application API (DRMAA) has been proposed to aid the rapid development and distribution of these applications across the grid. In this paper we present the first implementation of the DRMAA standard on a Globus-based testbed, and show its suitability to express typical scientific applications. As a case of study, we will consider in this paper the implementation of the NAS grid benchmarks with DRMAA. The DRMAA routines are supported by the functionality offered by the GridWay framework, that provides the runtime mechanisms needed for transparently executing jobs on a dynamic grid environment.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78374053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We investigate the problem of minimizing the makespan (resp. the sum of the completion time) for the multiprocessor scheduling problem in presence of hierarchical communications. We consider a model with two levels of communication: interprocessor and intercluster. The processors are grouped in connected clusters. We propose general non-approximability results for the case where all the tasks of the precedence graph have unit execution times, where the multiprocessor is composed of an unrestricted number of machines with /spl lscr/ /spl ges/ 4 identical processors each.
{"title":"General non-approximability results in presence of hierarchical communications","authors":"R. Giroudeau, J. König","doi":"10.1109/ISPDC.2004.27","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.27","url":null,"abstract":"We investigate the problem of minimizing the makespan (resp. the sum of the completion time) for the multiprocessor scheduling problem in presence of hierarchical communications. We consider a model with two levels of communication: interprocessor and intercluster. The processors are grouped in connected clusters. We propose general non-approximability results for the case where all the tasks of the precedence graph have unit execution times, where the multiprocessor is composed of an unrestricted number of machines with /spl lscr/ /spl ges/ 4 identical processors each.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84989158","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A new program graph structuring algorithm for dynamically reconfigurable multi-processor systems based on the look-ahead dynamic connection reconfiguration is presented. This architectural model enables elimination of connection reconfiguration time overheads. It consists in preparing link connections in advance in redundant connection switches in parallel with program execution. An application program is partitioned into sections, which are executed using such connections. Parallel program scheduling in this environment incorporates graph partitioning problem. The new algorithm is based on list scheduling and a new iterative clustering heuristics for graph partitioning. The experimental results are presented, which compare performance of several graph partitioning heuristics for such environment.
{"title":"Program structuring algorithms for dynamically reconfigurable parallel systems based on redundant connection switches","authors":"E. Laskowski","doi":"10.1109/ISPDC.2004.43","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.43","url":null,"abstract":"A new program graph structuring algorithm for dynamically reconfigurable multi-processor systems based on the look-ahead dynamic connection reconfiguration is presented. This architectural model enables elimination of connection reconfiguration time overheads. It consists in preparing link connections in advance in redundant connection switches in parallel with program execution. An application program is partitioned into sections, which are executed using such connections. Parallel program scheduling in this environment incorporates graph partitioning problem. The new algorithm is based on list scheduling and a new iterative clustering heuristics for graph partitioning. The experimental results are presented, which compare performance of several graph partitioning heuristics for such environment.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82090515","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We investigate the feasibility of running parallel applications on heterogeneous clusters. The motivation for doing so is twofold. First, it is practical to be able to pull together existing machines to run a job that is too big for any one of them, especially if such jobs are run rarely. Second, in the event of an emergency, where a very large problem must be solved in a few days, it may not be feasible to purchase and install a new machine in time, and any existing machines will have to be brought to bear on the problem. We ran the Multi-zone versions of the NAS Parallel Benchmarks (NPB) on a cluster composed of two SGI Origin 2000 servers, and an Intel SMP Xeon server connected by Gigabit Ethernet. We report on the results and their implications for running parallel applications on heterogeneous clusters.
{"title":"Load balancing multi-zone applications on a heterogeneous cluster with multi-level parallelism","authors":"P. Wong, Haoqiang Jin, J. Becker","doi":"10.1109/ISPDC.2004.33","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.33","url":null,"abstract":"We investigate the feasibility of running parallel applications on heterogeneous clusters. The motivation for doing so is twofold. First, it is practical to be able to pull together existing machines to run a job that is too big for any one of them, especially if such jobs are run rarely. Second, in the event of an emergency, where a very large problem must be solved in a few days, it may not be feasible to purchase and install a new machine in time, and any existing machines will have to be brought to bear on the problem. We ran the Multi-zone versions of the NAS Parallel Benchmarks (NPB) on a cluster composed of two SGI Origin 2000 servers, and an Intel SMP Xeon server connected by Gigabit Ethernet. We report on the results and their implications for running parallel applications on heterogeneous clusters.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89926737","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper describes a method of parallelisation of the popular Nelder-Mead simplex optimization algorithms that can lead to enhanced performance on parallel and distributed computing resources. A reducing set of simplex vertices are used to derive search directions generally closely aligned with the local gradient. When tested on a range of problems drawn from real-world applications in science and engineering, this reducing set concurrent simplex (RSCS) variant of the Nelder-Mead algorithm compared favourably with the original algorithm, and also with the inherently parallel multidirectional search algorithm (MDS). All algorithms were implemented and tested in a general-purpose, grid-enabled optimization toolset.
{"title":"RSCS: a parallel simplex algorithm for the Nimrod/O optimization toolset","authors":"A. Lewis, D. Abramson, T. Peachey","doi":"10.1109/ISPDC.2004.44","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.44","url":null,"abstract":"This paper describes a method of parallelisation of the popular Nelder-Mead simplex optimization algorithms that can lead to enhanced performance on parallel and distributed computing resources. A reducing set of simplex vertices are used to derive search directions generally closely aligned with the local gradient. When tested on a range of problems drawn from real-world applications in science and engineering, this reducing set concurrent simplex (RSCS) variant of the Nelder-Mead algorithm compared favourably with the original algorithm, and also with the inherently parallel multidirectional search algorithm (MDS). All algorithms were implemented and tested in a general-purpose, grid-enabled optimization toolset.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75243787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Peer to peer and grid systems provide attractive middlewares to solve large numerical problems. The development, deployment and execution of applications using those middlewares suffer from the lack of well-adapted advanced tools. There is not any available solution to use the same application on two distinct middlewares. Our article presents the YML Framework which provides supporting tools to design and execute portable parallel applications over large scale peer to peer and grid middlewares. The YML Framework defines a new parallel programming language called YvetteML which is composed of a graph language and a component model. We evaluate the performance of our framework with a simple numerical application using XtremWeb as a middleware.
{"title":"A peer to peer computing framework: design and performance evaluation of YML","authors":"O. Delannoy, S. Petiton","doi":"10.1109/ISPDC.2004.7","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.7","url":null,"abstract":"Peer to peer and grid systems provide attractive middlewares to solve large numerical problems. The development, deployment and execution of applications using those middlewares suffer from the lack of well-adapted advanced tools. There is not any available solution to use the same application on two distinct middlewares. Our article presents the YML Framework which provides supporting tools to design and execute portable parallel applications over large scale peer to peer and grid middlewares. The YML Framework defines a new parallel programming language called YvetteML which is composed of a graph language and a component model. We evaluate the performance of our framework with a simple numerical application using XtremWeb as a middleware.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87082265","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The paper presents a new architecture for systems based on run-time reconfigured shared memory processor clusters meant for implementation using network on chip technology. Clusters constitute local data exchange sub-networks, which dynamically connect processors with shared memory modules. The sub-networks enable exposure of data from one processor's data cache for reading by other processors to their data caches. This inter-processor data exchange paradigm, called "communication on the fly", enables direct communication between processor data caches. Dual-ported data caches are assumed to enable parallel reading and writing data between the caches and memory modules. In the proposed architecture, programs are executed according to a cache-controlled macro data flow execution model. Computational tasks are so defined, as to eliminate re-loading of data caches during task execution. A special program macro-data flow graph representation enables modeling of program behaviour for different architectural and program structure assumptions. Simulation results of symbolic execution of program graphs of matrix multiplication are presented in the paper. They show suitability of the proposed architecture for very fine grain parallel computations.
{"title":"Dynamic SMP clusters with communication on the fly in NoC technology for very fine grain computations","authors":"M. Tudruj, L. Masko","doi":"10.1109/ISPDC.2004.20","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.20","url":null,"abstract":"The paper presents a new architecture for systems based on run-time reconfigured shared memory processor clusters meant for implementation using network on chip technology. Clusters constitute local data exchange sub-networks, which dynamically connect processors with shared memory modules. The sub-networks enable exposure of data from one processor's data cache for reading by other processors to their data caches. This inter-processor data exchange paradigm, called \"communication on the fly\", enables direct communication between processor data caches. Dual-ported data caches are assumed to enable parallel reading and writing data between the caches and memory modules. In the proposed architecture, programs are executed according to a cache-controlled macro data flow execution model. Computational tasks are so defined, as to eliminate re-loading of data caches during task execution. A special program macro-data flow graph representation enables modeling of program behaviour for different architectural and program structure assumptions. Simulation results of symbolic execution of program graphs of matrix multiplication are presented in the paper. They show suitability of the proposed architecture for very fine grain parallel computations.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84975643","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Task scheduling is in general an NP-complete problem. For this reason a huge number of heuristics have been presented in the literature to obtain near optimal schedules. These heuristics mainly target homogeneous computing systems, while a few of them target heterogeneous systems. The heterogeneous heuristics presented so far target computing machines with different capabilities, while almost none of them handle heterogeneous communication systems. This paper presents a novel task scheduling algorithm called the heterogeneous critical tasks reverse duplicator (HCTRD), which targets both heterogeneous computation and communication systems. The algorithm is based on list-scheduling and task-duplication in a bounded number of machines, and aims to achieve high performance and near lower bound complexity.
{"title":"A near lower-bound complexity algorithm for compile-time task-scheduling in heterogeneous computing systems","authors":"Tarek Hagras, J. Janecek","doi":"10.1109/ISPDC.2004.3","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.3","url":null,"abstract":"Task scheduling is in general an NP-complete problem. For this reason a huge number of heuristics have been presented in the literature to obtain near optimal schedules. These heuristics mainly target homogeneous computing systems, while a few of them target heterogeneous systems. The heterogeneous heuristics presented so far target computing machines with different capabilities, while almost none of them handle heterogeneous communication systems. This paper presents a novel task scheduling algorithm called the heterogeneous critical tasks reverse duplicator (HCTRD), which targets both heterogeneous computation and communication systems. The algorithm is based on list-scheduling and task-duplication in a bounded number of machines, and aims to achieve high performance and near lower bound complexity.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90022556","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
R. Byrom, B. Coghlan, Andrew W. Cooke, Roney Cordenonsi, L. Cornwall, A. Datta, Abdeslem Djaoui, L. Field, S. Fisher, Steve Hicks, S. Kenny, J. Magowan, W. Nutt, D. O'Callaghan, M. Oevers, N. Podhorszki, John Ryan, Manish Soni, P. Taylor, A. Wilson, Xiaomei Zhu
We describe how the R-GMA (Relational Grid Monitoring Architecture) can be used to allow for instrument monitoring in a Grid environment. The R-GMA has been developed within the European DataGrid Project (EDG) as a Grid Information and Monitoring System. It is based on the Grid Monitoring Architecture (GMA) from the Global Grid Forum (GGF), which is a simple Consumer-Producer model. The special strength of this implementation comes from the power of the relational model. It offers a global view of the information as if each Virtual Organisation had one large relational database. It provides a number of different Producer types with different characteristics; for example some support streaming of information. We describe the R-GMA component that allows for instrument monitoring, the CanonicalProducer. We also describe an example use of this approach in the European CrossGrid project, SANTA-G, a network monitoring tool.
{"title":"The CanonicalProducer: an instrument monitoring component of the Relational Grid Monitoring Architecture (R-GMA)","authors":"R. Byrom, B. Coghlan, Andrew W. Cooke, Roney Cordenonsi, L. Cornwall, A. Datta, Abdeslem Djaoui, L. Field, S. Fisher, Steve Hicks, S. Kenny, J. Magowan, W. Nutt, D. O'Callaghan, M. Oevers, N. Podhorszki, John Ryan, Manish Soni, P. Taylor, A. Wilson, Xiaomei Zhu","doi":"10.1109/ISPDC.2004.50","DOIUrl":"https://doi.org/10.1109/ISPDC.2004.50","url":null,"abstract":"We describe how the R-GMA (Relational Grid Monitoring Architecture) can be used to allow for instrument monitoring in a Grid environment. The R-GMA has been developed within the European DataGrid Project (EDG) as a Grid Information and Monitoring System. It is based on the Grid Monitoring Architecture (GMA) from the Global Grid Forum (GGF), which is a simple Consumer-Producer model. The special strength of this implementation comes from the power of the relational model. It offers a global view of the information as if each Virtual Organisation had one large relational database. It provides a number of different Producer types with different characteristics; for example some support streaming of information. We describe the R-GMA component that allows for instrument monitoring, the CanonicalProducer. We also describe an example use of this approach in the European CrossGrid project, SANTA-G, a network monitoring tool.","PeriodicalId":62714,"journal":{"name":"骈文研究","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2004-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74251626","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：