Resource optimization in a LAN environment using SMIG-shared memory integrated with grid

Abstract

From cluster computing to Grid computing, current computation technologies have focused more on collaboration, data sharing, cycle stealing modes of interaction among dynamic and geographically distributed organizations. During the literature survey of the doctoral research the focus was on two computing paradigms: Grid computing and Distributed Shared Memory. There has been extensive research in individual implementations of both DSM and Grids but very few combined implementations. Hence in this paper there is focus on implementing and testing SMIG (Shared Memory Integrated with Grid) a model which integrates both paradigms. The paper discusses the design principles and architecture of SMIG, followed by explanation of the installation and execution process. Next the graphs are plotted and results are analyzed for SMIG execution for various applications like Matrix multiplication, Merge sort and Bucket sort. Based on the analysis it was inferred that SMIG model is feasible, can exploit existing underutilized resources and assist in resource optimization or carrying out high end computations. In future it is proposed to execute the same applications in a JUMP DSM environment and compare it with SMIG results.