A Cost-Driven Approach to Caching-as-a-Service in Cloud-Based 5G Mobile Networks

The exploding volumes of mobile video traffic call for deploying content caches inside mobile operator networks. Within-network caching, users’ requests for popular content can be served from a content cache deployed at mobile gateways in vicinity tothe end user. This inherently reduces the load on the content servers and the backbone of operator’s network. In light of the increasingtrend in virtualization of network functions, we propose a cost-effective caching as a service (CaaS) framework for virtual video cachingin 5G mobile networks. In order to evaluate the pros and cons of our CaaS approach, we formulate two virtual caching problems,namely maximum return on investment (MRI) and maximum offloaded traffic (MOT). MRI aims at maximizing return on cachinginvestment by finding the best trade-off between the cost of cache storage and bandwidth savings from caching video contents in themobile network operator (MNO)’s cloud. Likewise, MOT aims to maximize the traffic offloaded from the MNO’s core and backhaul withingiven budget constraints. More specifically, taking the popularity and size of video contents into account, MRI and MOT aim to find theoptimal caching tables which maximize the ratio of transmission bandwidth cost to storage cost and the offloaded traffic for a givenbudget, respectively. We reduce the complexity of the proposed problem formulated as a binary-integer programming (BIP) by usingcanonical duality theory (CDT). Experimental results obtained using the invasive weed optimization (IWO) have shown significantperformance enhancement of the proposed system in terms of return on investment, quality, offloaded traffic and storage efficiency.