Popular Content Caching in 5G and Beyond Mobile Networks

The 5G and beyond network system aims to satisfy the widest range of access and connectivity requirements in the mobile and wireless communications. In addition, it aims at providing an extremely flexible platform offering ubiquitous connectivity for any kind of device and service in the globe. The 5G wireless capability includes massive system capacity, very high data rate everywhere, very low latency, ultra-high reliability and availability, very low operational cost and energy consumption, and energy-efficient networks.

In the current heterogeneous cellular network (HCN), a small percentage of the mobile contents becomes highly popular, for a time period, and attracts frequent requests from many users, thus producing a large portion of the mobile data traffic in the network. Hence, proper management of such popular contents, e.g., deploying mobile data caching, based on content attributes, such as popularity, ephemerality or popularity span, and size has the potential for big network performance improvement.

A simple architecture of content caching in mobile networks

Caching of popular content is considered an opportunistic technology used to ease content availability in 5G and beyond mobile networks. Caching such contents locally, at the cache-enabled radio access network edge, brings them closer to the user. We assume that popular content is cached at predefined points at the network edge, referred to as mobile helpers (MH), which also serve the user requests.  This caching avoids recurrent fetching of contents from the backhaul network for each request and reduces traffic congestion and network response time which becomes even zero when the requested content is already cached at the user terminal (UT) itself. As a result, popular content caching has a significant role in achieving the expected quality of experience (QoE) to the user of the mobile network.

Popular content caching involves two main phases: content placement and content delivery. Several content transmission techniques have been proposed to enhance the efficiency of content placement and content delivery in the network, including channel coding, channel estimation, and limit resolution techniques.

Content Placement refers to storing the mobile contents from the content delivery network (CDN) or mobile backhaul network to the cache-enabled MHs at the mobile network edge, such as macrocell base stations (BS), microcell small base stations (SBS), and UTs. In this regard, several strategies have been proposed for optimizing content placement, from the point of view of both the information theory and physical layer management. An optimal content placement strategy enables the network to store the most popular and highly prioritized contents, as far as the cache size of the MH allows, so that users can be served from the cached contents. This placement scheme is done during the off-peak time period and is referred to as an off-line content placement. When the request is for a deadline-critical service, such as online streaming, content caching is immediately done as the request comes from the user; we will refer to such an immediate caching as online content placement. Having a prioritization strategy among the two placement cases leads to better backhaul congestion offloading. If the request comes from the caching device itself, then no extra caching is needed.

Content Delivery refers to transferring the required content from the MH to the UT (i.e., the effective decoding of popular contents at the UT). As the request is created, the content is served immediately if it is already cached in the UT’s own library; otherwise, the content is fetched from the appropriate MH library and is delivered to the UT. The content delivery, hence, depends on transmission media related factors and the efficiency of information decoding algorithms.

The design of content caching techniques should consider several factors, including the spatial flexibility of contents and UTs, the ephemerality of content popularity, the correlation of requests, the priority level of the requests, and user security. In addition, due to the limited storage capacity, a real-time updating strategy is required to keep only relevant contents in the cache library.

Fortunately, the advances in memory technology, i.e., caching capacity of current end-user devices, renders the proposed caching techniques practical in the emerging 5G mobile networks. The network performance improvement attained by the proposed caching techniques can be further enhanced by the deployment of complementary advanced techniques, such as device-to-device (D2D) communication, network clustering, cooperative and coordinated caching transmission medium control, and coded information transmission. In SPOTLIGHT, the research work on content caching focuses on the development of caching strategies with increased content delivery efficiency.

References:

1. Juan Wen, Kaibin Huang, Sheng Yang, and Victor O. K. Li, “Cache-Enabled Heterogeneous Cellular Networks: Optimal Tier-Level Content Placement”, IEEE Transactions on Wireless Communications, Vol.16, No.9, Sept. 2017.
2. Mehrnaz Afshang, Harpreet S. Dhillon, Peter Han Joo Chong, “Fundamentals of Cluster-Centric Content Placement in Cache-Enabled Device-to-Device Networks”, IEEE Transactions on Communications, Vol. 64, No. 6, June 2016.
3. T. Hoang, D. Niyato, D. N. Nguyen, E. Dutkiewicz, P. Wang, and Z. Han, “A Dynamic Edge Caching Framework for Mobile 5G Networks,” IEEE Wireless Communications, pp. 1–9, 2018