Title :
MEFTTD: A Parallel Awareness Scheduler
Author :
Zheng Yuan ; Bo Li
Author_Institution :
Dept. of Comput. Sci., Beihang Univ., Beijing, China
Abstract :
This paper presents a scheduling algorithm on cluster, named modified earliest finish time with task description (MEFTTD). Unlike other algorithms, MEFTTD needs programmers to describe the attribute of their parallel program. Then, MEFTTD uses these descriptions to optimize scheduling strategy. In this paper, round-robin algorithm (R-R), Heterogeneous Earliest-Finish-Time algorithm (HEFT) and MEFTTD are compared. The overall result demonstrates that MEFTTD can improve scheduling performance dramatically.
Keywords :
parallel programming; scheduling; HEFT algorithm; MEFTTD algorithm; R-R algorithm; heterogeneous earliest-finish-time algorithm; modified earliest finish time with task description; parallel awareness scheduler; parallel programming; round-robin algorithm; scheduling algorithm; scheduling performance; scheduling strategy; Cloud computing; Clustering algorithms; Delays; Schedules; Scheduling; Scheduling algorithms; cluster scheduling; parallel awareness; tasks description;
Conference_Titel :
Cloud and Service Computing (CSC), 2013 International Conference on
Conference_Location :
Beijing
DOI :
10.1109/CSC.2013.31