Coded caching (CC) schemes exploit the cumulative cache memory of the users and simple linear coding to turn unicast traffic (individual file requests) into a multicast transmission. For the originally proposed $K$-user single-server/single shared link network model, CC yields an $O(K)$ gain with respect to conventional uncoded caching with the same per-user memory. While several information-theoretic optimality results for a variety of problems and carefully crafted network topologies have been proved, the gains and suitability of CC for practical scenarios such as content streaming over existing wireless networks have not yet been fully demonstrated. In this work, we consider CC for on-demand video streaming over WLANs where multiple users are served simultaneously by multiple spatially distributed access points (AP). Users sequentially request video ``chunks". The CC scheme operates above the IP layer, leaving the underlying standard physical layer and MAC layer untouched. The cache placement is completely asynchronous and decentralized, and the users are placed at random over the network coverage area. For such a system, we consider the region of achievable long-term average delivery rate (defined as the number of video chunks delivered per unit of time) and study the per-user rate distribution under proportional fairness scheduling. We also consider reduced complexity scheduling strategies and compare them with standard state-of-the-art techniques such as conventional (uncoded) caching and collision avoidance by allocating APs on different sub-channels (i.e., frequency reuse).

翻译：编码缓存（CC）方案利用用户的累积缓存内存和简单线性编码，将单播流量（个体文件请求）转化为多播传输。对于最初提出的$K$用户单服务器/单共享链路网络模型，CC相较于同等每用户内存的传统非编码缓存，实现了$O(K)$倍的增益。虽然针对多种问题和精心设计的网络拓扑已证明若干信息论最优性结果，但CC在现有无线网络（如内容流传输）等实际场景中的增益和适用性尚未得到充分验证。本文研究无线局域网（WLAN）中按需视频流传输的编码缓存，其中多个用户同时由多个空间分布的接入点（AP）提供服务。用户按顺序请求视频“块”。CC方案运行于IP层之上，不改变底层标准物理层和MAC层。缓存放置完全异步且分散，用户随机分布于网络覆盖区域。针对该系统，我们考虑可实现的长期平均传输速率（定义为每秒交付的视频块数量）区域，并研究比例公平调度下的每用户速率分布。此外，我们还考察了简化复杂度的调度策略，并将其与采用不同子信道（即频率复用）分配AP以避免冲突的标准先进技术（如传统非编码缓存）进行对比。