By Topic

A Heuristic Algorithm for Scheduling on Grid Computing Environment

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)
Jing Wang ; Sch. of Comput. Sci. & Inf. Eng., Hefei Univ. of Technol., Hefei, China ; Gongqing Wu ; Bin Zhang ; Xuegang Hu

With the conglomeration of large-scale heterogeneous systems, the grid computing environment makes the whole network into a powerful and reliable resource available nearly everywhere. Resource scheduling is a fundamental issue in grid computing. For this NP-hard problem, we take into account of the geographic distribution of resources and the requirement of job entity in the scheduling algorithm. To do so, we first consider the parameters of job entity and resource entity. Then the key characteristics as release time, processing time and delivery time determine the rules about the scheduling. We present HF (Harder First) strategy and DF (Larger Distance First) strategy. Let the H value denotes the sum of release time, length and delivery time of the job, the job with a higher H value is considered to be harder and should be assigned to a faster resource according to the HF strategy. Secondly, when the number of jobs is larger than the number of resources, the DF strategy makes sure that the job with a higher difference (distance) between the delivery time and the release time should be processed first. Based on the stated strategies, we provide a heuristic algorithm HFFP (Harder First Faster Prior) for resource scheduling on the grid computing environment. The experiment data of jobs scale from 10k to 80k, while the number of resources ranges from 2 to 6. The algorithm performance is demonstrated by simulation on the platform of GridSim. Our experiment results show that the algorithm HFFP can minimize the completion time of jobs especially when the number of jobs is much larger than the number of resources. By comparing our algorithm with classical scheduling algorithm as Min-min algorithm, we can see that our algorithm can assign the jobs to the resources reasonably from the criteria of make span. To better compare the performance of our algorithm with Max-min, we do some medication to the traditional Max-min algorithm and presents Max-min-L (Max-min-Local). Max-m- n-L chooses the local maximization instead of overall maximization, suitable for jobs with similar length. By comparing experiments with Max-min-L and Min-min, we can still get that our algorithm is better than Min-min and Max-min-L by the metrics of make span.

Published in:

ChinaGrid Annual Conference (ChinaGrid), 2012 Seventh

Date of Conference:

20-23 Sept. 2012