Dynamic Equilibrium Replica Location Algorithms in Data Grid

Replica location is one of the key issues of data management in grid environment. Existing replica location services employ statistical location information to locate replica, lack initiative detecting mechanism for update, and equalize location information at large cost. In this paper, aimed at existing problems, combined with the characteristics of data gird, a dynamic equilibrium replica location algorithm (DERLS) was proposed. The idea underlying our DERLS algorithms consists of three parts: (1) Replica location service locates physical replica according to the density of pheromone by artificial ant spread. (2) A dynamic equilibrium technique is proposed to prevent positive feedback of basic ant algorithms. (3) Update strategy based on improved ant algorithms is used to detect new replica or resume fault replica. Experimental results show the availability and validity of the replica location in detail.

[1]  Ian Foster,et al.  A Decentralized, Adaptive, Replica Location Service , 2002 .

[2]  Kai Lu,et al.  Decentralized adaptive replica location mechanism in large-scale distributed systems , 2003, Proceedings of the Fourth International Conference on Parallel and Distributed Computing, Applications and Technologies.

[3]  Kurt Stockinger,et al.  Simulation of Dynamic Grid Replication Strategies in OptorSim , 2002, GRID.

[4]  Chen Ling,et al.  Adaptive Exchanging Strategies in Parallel Ant Colony Algorithm , 2007 .

[5]  Ian T. Foster,et al.  The data grid: Towards an architecture for the distributed management and analysis of large scientific datasets , 2000, J. Netw. Comput. Appl..

[6]  Ian T. Foster,et al.  A decentralized, adaptive replica location mechanism , 2002, Proceedings 11th IEEE International Symposium on High Performance Distributed Computing.

[7]  Krishna P. Gummadi,et al.  Measuring and analyzing the characteristics of Napster and Gnutella hosts , 2003, Multimedia Systems.