A decentralized strategy for resource allocation

Abstract

This paper presents an approach for resource allocation in a grid based on "spatial computing" concepts. We model a grid using a flat architecture consisting of nodes connected by an overlay network. The 2D spatial distribution of the nodes of the grid, together with the quantity of resource available in each node, forms a 3D surface, where valleys correspond to nodes with a large quantity of available resource. We propose an algorithm for resource allocation that is based on surfing such a 3D surface, in search for the deepest valley (global minimum). The algorithm, which aims at fairly distributing among nodes the quantity of leased resource, is based on some heuristics that mimic the laws of kinematics. Experimental results show the effectiveness of the algorithm.