Title :
A dynamic sender-initiated load sharing algorithm for cluster with reduced polling
Author :
Sandhya, K.V. ; Raju, G.
Author_Institution :
Dept. of Comput. Sci., Karpagam Univ., Coimbatore, India
Abstract :
It is very important that algorithms used in cluster do the work in the shortest time possible. Only efficient communication between the cluster nodes can improve the performance of a cluster. Considering this aspect, redundancy of work can be avoided if nodes share information. The proposed algorithm is based on the idea of sharing of information already acquired by nodes so that less work is done by individual nodes for load distribution. The algorithm focuses on improving the performance of a cluster computing environment, by sharing cluster statistics information between individual nodes and thus reduces the time taken to find valid receivers. The implementation of the algorithm using a simulator is ongoing work.
Keywords :
resource allocation; workstation clusters; cluster nodes; efficient communication; load distribution; reduced polling; sender-initiated load sharing; Algorithm design and analysis; Clustering algorithms; Heuristic algorithms; Load management; Peer to peer computing; Receivers; Vectors; cluster computing; load sharing; performance;
Conference_Titel :
Recent Trends in Information Systems (ReTIS), 2011 International Conference on
Conference_Location :
Kolkata
Print_ISBN :
978-1-4577-0790-2
DOI :
10.1109/ReTIS.2011.6146858