Combining software agents and grid middleware
GPC'07 Proceedings of the 2nd international conference on Advances in grid and pervasive computing
Data mining on desktop grid platforms
PPAM'07 Proceedings of the 7th international conference on Parallel processing and applied mathematics
Extremal optimization approach applied to initial mapping of distributed java programs
EuroPar'10 Proceedings of the 16th international Euro-Par conference on Parallel processing: Part I
Extremal optimization applied to task scheduling of distributed Java programs
EvoApplications'11 Proceedings of the 2011 international conference on Applications of evolutionary computation - Volume Part II
Distributed java programs initial mapping based on extremal optimization
PARA'10 Proceedings of the 10th international conference on Applied Parallel and Scientific Computing - Volume Part I
Hi-index | 0.00 |
In Java computing context, ADAJ (Adaptive Distributed Application in Java) provides a both execution and programming system for irregular distributed applications. From the execution point of view, load balancing is an important issue and requires special information policies.Traditionnally only the computer load is taken into account. Consequently, the load balancing is made by being unaware of the relations between objects. The objects are moved of acomputer on the other without knowing the consequences in term of communication. To avoid that problem we introduce in that paper two mechanims: a relation observation mechanism of distributed objects and a computer load observation. The dynamical observation information provided by these mechanisms will be able to allow a general load balancing mechanism to have an intelligent adaptative object redistribution strategy. The proposed observation tools are entirely designed in Java. Some experiment results whichconcerns relation observations and overhead measurements are presented.