Title :
A grid execution environment similarity metric
Author :
Wang, Wei ; Fang, Bin-Xing
Author_Institution :
Dept. of Comput. Sci., Harbin Inst. of Technol., China
Abstract :
A program´s behavior is always tightly related with the computing environment where it runs, especially for parallel programs running in grid. Debugging tries to locate the reason for incorrect program behavior by analyzing the states, which happen during a program´s execution. In order to allow cyclic debugging, re-execution mechanisms are used most frequently. The dynamic characteristics of grid will hinder re-execution to work, because the behavior of a program will be various according to the current configuration of execution environment. By modeling grid environments with graphics, this paper proposes the metric to evaluate the similarity between grid applications execution environments. The GA based resource selection algorithm is presented to construct similar execution environments in grids. A serial of experiments shows that the similarity metrics is appropriate.
Keywords :
genetic algorithms; grid computing; program debugging; program diagnostics; resource allocation; computing environment; genetic algorithm; grid applications; grid environments model; grid execution environment; parallel debugging; parallel programs; resource management; resource selection algorithm; similarity metric; Concurrent computing; Costs; Debugging; Distributed computing; Graphics; Grid computing; Machine learning algorithms; Resource management; Runtime; Software testing; Grid; computing environment; genetic algorithm; parallel debugging; resource management; similarity metric;
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
DOI :
10.1109/ICMLC.2005.1527458