Title :
Algorithmic mechanism design for load balancing in distributed systems
Author :
Grosu, Daniel ; Chronopoulos, Anthony T.
Author_Institution :
Dept. of Comput. Sci., Texas Univ., San Antonio, TX, USA
Abstract :
Computational grids are large scale computing systems composed of geographically distributed resources (computers, storage etc.) owned by self interested agents or organizations. These agents may manipulate the resource allocation algorithm for their own benefit and their selfish behavior may lead to severe performance degradation and poor efficiency. In this paper we investigate the problem of designing protocols for resource allocation involving selfish agents. Solving this kind of problem is the object of mechanism design theory. Using this theory we design a truthful mechanism for solving the static load balancing problem in heterogeneous distributed systems. We prove that by using the optimal allocation algorithm the output function admits a truthful payment scheme satisfying voluntary participation. We derive a protocol that implements our mechanism and present experiments to show its effectiveness.
Keywords :
distributed algorithms; protocols; resource allocation; workstation clusters; algorithmic mechanism design; computational grids; geographically distributed resources; heterogeneous distributed systems; large scale computing system; optimal allocation algorithm; output function; protocols; resource allocation algorithm; self interested agents; selfish agents; static load balancing problem; truthful payment scheme; voluntary participation; Algorithm design and analysis; Computer science; Cost accounting; Degradation; Distributed computing; Grid computing; Large-scale systems; Load management; Protocols; Resource management;
Conference_Titel :
Cluster Computing, 2002. Proceedings. 2002 IEEE International Conference on
Print_ISBN :
0-7695-2066-9
DOI :
10.1109/CLUSTR.2002.1137780