PROJECT TITLE :

Adaptive Caching Networks With Optimality Guarantees - 2018

ABSTRACT:

We have a tendency to study the optimal placement of content over a network of caches, a problem naturally arising in several networking applications. Given a demand of content request rates and ways followed, we tend to would like to see the content placement that maximizes the expected caching gain, i.e., the reduction of routing prices because of intermediate caching. The offline version of this drawback is NP-arduous and, normally, the demand and topology might be a priori unknown. Hence, a distributed, adaptive approximation algorithm for inserting contents into caches is desired. We have a tendency to show that path replication, a straightforward algorithm frequently encountered in literature, will be arbitrarily suboptimal when combined with traditional eviction policies. We propose a distributed, adaptive algorithm that performs stochastic gradient ascent on a concave relaxation of the expected caching gain, and constructs a probabilistic content placement among a one-one/e factor from the optimal, in expectation. Motivated by our analysis, we have a tendency to also propose a novel greedy eviction policy to be used with path replication, and show through numerical evaluations that both algorithms considerably outperform path replication with traditional eviction policies over a broad array of network topologies.