Scheduling of Workflows in Grid Computingwith Probabilistic Tabu Search
R. Joshua Samuel Raj
CSE, VV College of EngineeringTirunelveli, India email@example.com
Dr. V. Vasudevan
Prof. & Head/IT, Kalasalingam UniversitySrivilliputur, Indiadrvvmca@yahoo.com
In Grid Environment the number of resourcesand tasks to be scheduled is usually variable anddynamic in nature. This characteristic emphasizes thescheduling approach as a complex optimizationproblem. Scheduling is a key issue which must be solvedin grid computing study and a better scheduling schemecan greatly improve the efficiency.The objective of thispaper is to explore the Probabilistic Tabu Search topromote compute intensive grid applications tomaximize the Job Completion Ratio and minimizelateness in job completion based on the comprehensiveunderstanding of the challenges and the state of the artof current research. Experimental results demonstratethe effectiveness and robustness of the proposedalgorithm. Further the comparative evaluation withother scheduling algorithms such as First Come FirstServe (FCFS), Last Come First Serve (LCFS), EarliestDeadline First (EDF) and Tabu Search are plotted.
Key words: grid computing, workflow, Tabu Search, scheduling problem, Probabilistic Tabu Search
INTRODUCTIONGrid Computing a pioneer technique inharnessing the geographically dislocated computerpower has changed the perception on the utility andavailability of the computer power, which has carveda new technology that openly ventures andamalgamates an infinite number of computingdevices into any grid environment, augmenting to thecomputing capability and providing resolutions to thevarious tasks within the operational grid environmentbasically by enabling, sharing, selection andaggregation of geographically distributedautonomous resources dynamically at runtime,depending on their availability, capability,performance and cost, thereby shifting the focus tocollaborative environments, federating services andexchanging transactions in a mutual manner to shareresources and thereby achieve common goals toenhance productivity and speed up progress in muchthe same way that the Internet did in yesterdayseconomy, paving the way for numerous researchefforts in grid scheduling mechanismsGrid Computing is our greatest hope fordelivering computing as utility to homes and offices.Many large scale applications such as scientific,engineering and business problems (Hai
2002) are solved effectively usingthe logical amalgamation of geographically dispersedGrid resources (Bernan
2002). Grid computing,analogous to the pervasive electrical power grid,enables resource sharing and cooperative work among distributed computational sites.In grid environment, applications are often describedas workflows. A workflow is composed of atomictasks that are processed in specific order to fulfill acomplicated goal. Generally, grid workflows requirehuge intensive computing and process larger data,compared with traditional workflows. Therefore, theperformance of grid workflows becomes a criticalissue of the workflow management systems. One of the most challenging problems is to map each task to a corresponding service instance to achieve thecustomers’ quality of service (QoS) requirements aswell as to accomplish high performance of theworkflow. This problem is found to be NP-complete.During the course of grid scheduling there are manychallenges that require the simultaneous optimizationof several incommensurable and competingobjectives.
Unpredictable challenges in Grid resources
Inevitability to multiple resource types forcompleting a job
Necessitate for a parallel or concurrentexecution of tasks in any workflows.Under the OSGA, the workflow schedulerhas to balance several QoS requirements, includingmakespan and cost. Consequently, many traditionalworkflow scheduling algorithms, such asOpportunistic Load Balancing, Minimum CompletionTime, Min-min, Max-min and Duplex, are not
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 8, No. 4, July 2010314http://sites.google.com/site/ijcsis/ISSN 1947-5500