Assignment of DAGs Applications in Grid Computing Environment

The development of computer science especially Grids and networks is the concrete proof of technological advancement over the last thirty years. The complexities of mechanisms for analysis led us to elaborate ever more established models. The aggregation of resources has emerged immediately at the onset of networks, known as “union is power”, which gave birth to Grid Computing. These infrastructures used to provide computing capabilities virtually endless which is called “Grid Computing”. But among the problems encountered: the developments of an appropriate assignment policy for operations to be performed by a Grid. This is among the others to schedule a variety of applications composed of a set of dependent tasks and placing them, effectively, in minimizing the engine response time. This problem in the combinatorial analysis field is well-known as “NP-hard”, and becomes even more complex when tasks to assign contain precedence constraints. In this paper, we propose an assignment strategy in Grid computing environment, which incorporates dynamic heuristic assignment based on a clustered model (using a clustering algorithm). Our main goal is to assign tasks effectively reaching a minimum response time while respecting the constraints of dependency.