Research
Simulator for Resource Optimization of Job Scheduling in a Grid Framework
Traditionally, computer software’s has been written for serial computation. This software is to be run on a single computer with a single Central Processing Unit (CPU). A problem is broken into a discrete serial of instructions that executed in the exact order, one after another. Only one instruction can be executed at any moment of time on a single CPU. Parallel computing, on the other hand, is the simultaneous use of multiple computer resources to solve a computational problem. The program is to be run using multiple CPU’s. A problem is broken into discrete parts that can be solved concurrently and executed simultaneously on different CPU’s. The purpose of this proposed work is to develop a simulator using Java for the implementation of Job scheduling and shows that Parallel Execution is efficient with respect to serial execution in terms of time, speed and resources.
Sim_Dsc: Simulator for Optimizing the Performance of Disk Scheduling Algorithms
Disk scheduling involves a careful examination of pending requests to determine the most efficient way to service these requests. A disk scheduler examines the positional relationship among waiting requests, then reorders the queue so that the requests will be serviced with minimum seek. The purpose of the study is to obtain the best scheduling algorithm based on the seek time, rotation time and transfer time for moveable head disks. Keeping in view an attempt has been made to design a simulator for optimizing the performance of disk scheduling algorithms using Box-Muller transformation. The input for the simulator has been derived by using an algorithm for generating pseudo random numbers which follows box-muller transformations. Simulator takes access time which is generated using seek time, rotation time and transfer time, as the request of cylinder numbers, current position of read/write head as inputs. On the basis of these inputs, total head movement of each disk scheduling algorithm is calculated under various loads.
Handling of Congestion in Cluster Computing Environment Using Mobile Agent Approach
Computer networks have experienced an explosive growth over the past few years and with that growth have come severe congestion problems. Congestion must be prevented in order to maintain good network performance. In this paper, we proposed a cluster based framework to control congestion over network using mobile agent. The cluster implementation involves the designing of a server which manages the configuring, resetting of cluster. Our framework handles - the generation of application mobile code, its distribution to appropriate client, efficient handling of results, so generated and communicated by a number of client nodes and recording of execution time of application. The client node receives and executes the mobile code that defines the distributed job submitted by server and replies the results back. We have also the analyzed the performance of the developed system emphasizing the tradeoff between communication and computation overhead. The effectiveness of proposed framework is analyzed using JDK 1.5.
