In-network caching has evidently emerged as an indispensable core functionality of ICN. Of the various types of caching techniques, off-path caching turns out to be a potential technique. With a view to improve the performance of hash-based off-path caching, the aim of the paper is to ensure profitable, economic and selective placement of caches such that the final cache allocation tends to reduce simultaneously average retrieval delay and maximum internal link-stress. To achieve the same, we propose simple yet elegant heuristic algorithm to strategically identify and exclude the bad node-positions for placement of caches. More specifically, for a given topology, the algorithm computes cost (in terms of delay) for all the nodes and prepares a cost based descending ordered list of nodes. Further, iteratively, increasing number of nodes, in sequence, from the cost-based ordered list of nodes are debarred from caching thereby resulting in family of solutions. A solution represents a selective cache allocation map corresponding to the number of debarred nodes. Moreover, every iterative solution reduces average retrieval delay and seeks opportunistic reduction in maximum internal linkstress. Comparing the solutions graphically would enable network operators to finally select (as per desire) one cache allocation map for a given network topology for hash-based off-path caching. In a way, the selected cache allocation map for the considered topology implies that it is better to place the caches at the core of the network as compared to edge.
展开▼
