A novel approach to efficient resource allocation and scheduling in mobile grid

Abstract

In Grid computing or distributed computing, a central server is used to divide a great computational complexity into a group of autonomous workloads. Afterwards, these workloads are transmitted to a quantity of associated autonomous computing nodes which are encompassing own Central Processing Unit, storage and local memory. Then the workloads utilize their own computing resources to congregate the computational consequences. In this, the progression can take place by means of return to the parent server and download an innovative group of workloads can be downloaded. Generally, the computing resources are used to resolve a huge computational trouble, but they are costly and difficult. A commercial choice is made to distribute these workloads to a huge quantity of personal computers whose proprietors contribute to the redundant CPU time of their computers. According to the high workload and mobility of grid nodes, the level of overhead is increased and the resource reliability is reduced. In the present study, Mobility-Aware Resource Allocation and Scheduling (MARAS) Using Particle Swarm Optimization (PSO) in mobile grid is proposed to conquer these problems. Generally, a mobile monitoring agent is used to observe the mobility model and the movement direction of the node which is derived from the normal walk mobility representation. In this, the target and anticipated achievement times are predicted by this mobility. In case, if the augmented communication distance is bigger than a threshold value, then the remote communication bound tasks will be transferred to topologically closer nodes for enhanced resource allocation. At last, the PSO is employed for resource scheduling wherein the fitness function is calculated by the help of limitations like communication reliability and delay metric newline