FRDT: Footprint Resource Discovery Tree for grids

  • Authors:
  • Leyli Mohammad Khanli;Saeed Kargar

  • Affiliations:
  • Cs Department, University of Tabriz, Tabriz, Iran;Tabriz Azad University, Tabriz, Iran

  • Venue:
  • Future Generation Computer Systems
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Today, with the development of grid environments and an increased number of resources and geographical distribution, finding a resource discovery algorithm that can provide a user's required resources in a short time with low traffic, is an important task in resource discovery. Although recently developed approaches remove many of the defects of previous methods such as single points of failure and heavy loads as well as significantly improving the system efficiency, a user's query is again sent to unnecessary paths, so regarding the large number of nodes in grid environments, the system efficiency is reduced due to this additional traffic. In this paper, we propose a new algorithm and its use of a weighted tree for resource discovery. We use a bitmap where the number of bit positions is proportional to the number of attributes for each resource. Actually, in every node of the tree there will be a footprint of existence resources in children and the descendant of that node, so when the user's request reaches each node, if the requested resources are available in its children or the descendant, we can directly and without referring to extra nodes and unnecessary traffic reach the node which owns the required resource. We compare our algorithm with previous algorithms using simulations and results and show that the number of nodes visited in our resource discovery algorithm is less than that for other algorithms, and the difference would be significant with an increase in the number of nodes. Also the cost of update in our proposed algorithm is low.