SimBA: A Discrete Event Simulator for Performance Prediction of Volunteer Computing Projects
Proceedings of the 21st International Workshop on Principles of Advanced and Distributed Simulation
Simulation-based optimization of multiple-task GRID applications
Future Generation Computer Systems
An analytical model for performance evaluation in a computational grid
CHINA HPC '07 Proceedings of the 2007 Asian technology information program's (ATIP's) 3rd workshop on High performance computing in China: solution approaches to impediments for high performance computing
Defining and controlling the heterogeneity of a cluster: The Wrekavoc tool
Journal of Systems and Software
Wrekavoc: a tool for emulating heterogeneity
IPDPS'06 Proceedings of the 20th international conference on Parallel and distributed processing
Validation of SymGrid-services framework through event-based simulation
International Journal of Grid and Utility Computing
Performance oriented development and tuning of GRID applications
PARA'04 Proceedings of the 7th international conference on Applied Parallel Computing: state of the Art in Scientific Computing
Testing large-scale cloud management
IBM Journal of Research and Development
| Hi-index | 0.00 |
Improvements in networking and middleware technologyare enabling large-scale grids that aggregateresources over wide-area networks to support applicationsat unprecedented levels of scale and performance.Unfortunately, existing middleware and tools provide littleinformation to users as to the suitability of a given gridtopology for a specific grid application. Instead, users generallyuse ad-hoc performance models to evaluate mappingsof their applications to resource and network topologies.Grid application behavior alone is complex, andadding resource and network behavior makes the situationeven worse. As a result, users typically employ nearlyblind experimentation to find good deployments of their applicationsin each new grid environment. Only through actualdeployment and execution can a user discovers ifthe mapping was a good one. Further, even after findinga good configuration, there is no basis to determineif a much better configuration has been missed. This approachslows effective grid application development anddeployment.We present a richer methodology for evaluating gridsoftware and diverse grid environments based on the MicroGridgrid online simulator. With the MicroGrid, users,grid researchers, or grid operators can define and simulatearbitrary collections of resources and networks. This allowsstudy of an existing grid testbed under controlled conditionsor even to study the efficacy of higher performance environmentsthan are available today. Further, the MicroGridsupports direct execution of grid applications unchanged.These application can be written with MPI, C, C++, Perl,and/or Python and use the Globus middleware. This enablesdetailed and accurate study of application behavior. Thispaper presents: (1) the first validation of the MicroGrid forstudying whole-program performance of MPI Grid applicationsand (2) a demonstration of the MicroGrid as a toolfor predicting the performance of applications on a rangeof grid resources and novel network topologies.