The exploding volumes of mobile video traffic call for deploying content caches inside mobile operator networks. With in-network caching, users’ requests for popular content can be served from a content cache deployed at mobile gateways in vicinity to the end user. This inherently reduces the load on the content servers and the backbone of operator's network. In light of the increasing trend in virtualization of network functions, we propose a cost-effective caching as a service (CaaS) framework for virtual video caching in 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 caching investment by finding the best trade-off between the cost of cache storage and bandwidth savings from caching video contents in the mobile network operator (MNO)'s cloud. Likewise, MOT aims to maximize the traffic offloaded from the MNO's core and backhaul within given budget constraints. More specifically, taking the popularity and size of video contents into account, MRI and MOT aim to find the optimal caching tables which maximize the ratio of transmission bandwidth cost to storage cost and the offloaded traffic for a given budget, respectively. We reduce the complexity of the proposed problem formulated as a binary-integer programming (BIP) by using canonical duality theory (CDT). Experimental results obtained using the invasive weed optimization (IWO) have shown significant performance enhancement of the proposed system in terms of return on investment, quality, offloaded traffic, and storage efficiency.

中文翻译：

---

在基于云的 5G 移动网络中缓存即服务的成本驱动方法

激增的移动视频流量要求在移动运营商网络内部署内容缓存。通过网络内缓存，用户对流行内容的请求可以从部署在最终用户附近移动网关的内容缓存中得到满足。这从本质上减少了内容服务器和运营商网络主干的负载。鉴于网络功能虚拟化的增长趋势，我们提出了一种经济高效的缓存即服务 (CaaS) 框架，用于 5G 移动网络中的虚拟视频缓存。为了评估我们的 CaaS 方法的优缺点，我们制定了两个虚拟缓存问题，即最大投资回报 (MRI) 和最大卸载流量 (MOT)。MRI 旨在通过在缓存存储成本和在移动网络运营商 (MNO) 的云中缓存视频内容所节省的带宽之间找到最佳权衡，从而最大限度地提高缓存投资回报。同样，MOT 旨在在给定的预算限制内最大限度地减少从 MNO 的核心和回程分流的流量。更具体地说，考虑到视频内容的流行度和大小，MRI 和 MOT 旨在找到最佳缓存表，分别在给定预算下最大化传输带宽成本与存储成本和卸载流量的比率。我们通过使用规范对偶理论 (CDT) 降低了提出的问题的复杂性，该问题表示为二进制整数规划 (BIP)。