Data offloading in cache-enabled cross-haul networks

Abstract

This paper studies a cache-enabled cross-haul network (Ce-XHaul) infrastructure under the banner of 5G, aiming at offloading data requests in the backhaul networks by means of extremely restricting the provisioning of user requests in the fronthaul networks. The major motivation behind Ce-XHaul is to install a proper amount of cache storage in base stations (BSs) or in helper devices that are deployed between BSs and user devices (UEs), so as to store the frequently requested files from users, which sequentially reduce the requirement of access to the distant backhaul networks. Specifically, we first numerically study the network performance of Ce-XHaul in terms of latency, say the average transmission end-to-end (E2E) delay, under three network scenarios that differ with each other in the caching capabilities, named as no-caching (scenario 1), caching-at-BSs (scenario 2), and caching-at-helper nodes (scenario 3), respectively. Then, we dedicate on scenario 2 and investigate the impact of caching capability of BSs on the average transmission E2E delay by proposing two integer linear programming (ILP) models. The two ILP models are designed based on routing algorithms that differ in the efforts made to restrict the data provisioning in the fronthaul. Numerical results show that the more caching capability in the fronthaul, the less the latency, nonetheless, with higher CAPEX cost being induced due to deployment of the cache storage and the additional helper nodes.