Scheduling divisible load on Wireless Grid with communication Delay

Grid computing is an attractive and effective environment for High Performance parallel Computation. It is also the basis for cloud computing and has attracted many researchers to the field. Since scheduling drastically affects the performance of Grids, one can find in literature many articles that discuss scheduling in Grid computing Systems. In this paper we exploit the Divisible Load Theory (DLT) to model the scheduling of arbitrarily divisible load on the Grid. Most of the previous attempts do not consider the communication time. You can find some research in literature where communication time is well thought-out but not in dividing the load. So, the transfer input time of the load was not part of the model. In some work where the communication and computation time are considered, they failed to provide a closed form solution for the minimum execution time. In this paper we managed to alleviate the shortcoming of the previous work and we found a closed form solution for the minimum execution time of executing an arbitrarily divisible application on the Grid taking into consideration the communication time and the computation time. The execution time calculated by our analytical solution is the optimum (minimum). This work is a good starting point towards an analytical solution to such intractable problem.