Enhancing wireless networks with caching: Asymptotic laws, sustainability & trade-offs

We investigate on the sustainability of multihop wireless communications in the context of Information-Centric Networks, when content is replicated in caches over the network. The problem is cast in a flat wireless network for a given content popularity distribution and sized by three parameters, (i) the network size N, (ii) the content volume M and (iii) the cache capacity K per node. The objective is to select a joint replication and delivery scheme that minimizes the link traffic. Assuming the Zipf distribution about the content popularity, a law well established in the research on Internet traffic, we compute an order optimal solution, let the three size parameters jointly scale to infinity, and find the scaling laws about the link rates, ranging from ON down to o(1). Analyzing the derived laws, we determine the regimes that the network becomes sustainable subject to the scaling of the three network size parameters and the Zipf rank exponent, characterize the relative merit of network resources and identify the induced trade-offs about network expansion.