Jointly optimal routing and caching for arbitrary network topologies

Abstract

We study a problem of fundamental importance to ICNs, namely, minimizing routing costs by jointly optimizing caching and routing decisions over an arbitrary network topology. We consider both source routing and hop-by-hop routing settings. The respective offline problems are NP-hard. Nevertheless, we show that there exist polynomial time approximation algorithms producing solutions within a constant approximation from the optimal. We also produce distributed, adaptive algorithms with the same approximation guarantees. We simulate our adaptive algorithms over a broad array of different topologies. Our algorithms reduce routing costs by several orders of magnitude compared to prior art, including algorithms optimizing caching under fixed routing.

Joint Power Control and Caching for Transmission Delay Minimization in Wireless HetNets

“…We emphasize that our joint optimization framework is significantly different from the traditional approach which maximizes the caching gain only. This approach has been widely studied in the literature, such as in [2], [13], [24], and [52] and their follow-up works, where the link costs are fixed. This assumption is only true when the links are granted orthogonal frequencies and do not interfere, and the transmission powers are fixed, which is not the case in Het-Nets.…”

“…For example, the gain offered by local caching and broadcasting is characterized in [14]. Works also include the analysis of the scaling of the peruser throughput and collaboration distance [17], the wireless caching capacity region which is the closure of the set of all achievable caching traffic [19], as well as single-hop and device-to-device [14], [17], [20], [21], [22], and multi-hop caching networks [23], [24].…”

“…The parameter G vu is the channel power gain that includes only path loss, where we use the standard power loss propagation model, i.e., G vu = r −n vu given distance r vu between v and u, and the path loss exponent n > 2 [59]. The signal for request (i, p) over link (v, u) is decoded regarding all other signals as noise, for all (i, p) ∈ R and (v, u) ∈ E. Thus, in our model the transmission delays are coupled, in contrast to [2] and [24], because the decoding model captures the interference due to simultaneous wireless transmissions. Because the SINR analysis in ( 1) is for a single frequency band, the set of active nodes with nonzero transmission powers causes interference to the unintended receiver node.…”

“…where the last inequality also follows from the complementary slackness condition in (24). □ If D (i,p) (Y, S) is jointly convex in (Y, S), the above conditions are also sufficient when (25) holds for all v ∈ V .…”

Two-Layer Network Caching for Different Service Requirements

“…In the context of content centric networking, content placement schemes and request routing are tightly coupled. The relevant optimization problems typically consists maximization of a submodular functions subject to matroid constraints [29,30]. Such mathematical framework is also used in this paper.…”