[论文解读] Novel Inter-file Coded Placement and D2D Delivery for a Cache-aided Fog-RAN Architecture.

Kai Wan, Daniela Tuninetti+2 more|arXiv (Cornell University)|Nov 13, 2018

Caching and Content Delivery参考文献 37被引用 12

一句话总结

本文提出了一种新型的跨文件编码缓存放置方案与D2D缓存方案，适用于具有宏基站和小基站的缓存增强型雾无线接入网络，通过跨小基站通信实现高效的文件传输。该方案在最坏情况需求下达到常数因子范围内的最优性能，并可适应异构缓存大小与网络拓扑结构。

ABSTRACT

This paper considers a novel cache-aided fog Radio Access Network (Fog-RAN) architecture including a Macro-cell Base Station (MBS) and several Small-cell Base Stations (SBSs), serving users without caches. Some users, not in the reach of SBSs, are directly served by the MBS. Other users, in the reach of SBSs, are offloaded and receive information only from the SBSs through high throughput links. The SBSs have their own local caches. The MBS broadcasts packets to the SBSs and the directly served users. Subsequently, the SBSs can communicate among each other such that each SBS can obtain enough information to decode demanded files by its connected users. For the proposed Fog-RAN we study the memory-loads tradeoff for the worst-case demands. The main contribution of this paper is the design of a novel inter-file coded cache placement and a novel D2D caching scheme with shared-caches to handle the inter-SBS communication phase. The proposed scheme is proved to be order optimal within a constant factor when each SBS is connected to the same number of users. As by-product, the proposed inter-file coded placement can also be used in coded caching systems with heterogeneous cache sizes, where some users have no cache. The proposed coded placement depends on the network topology only by the number of SBSs and the number of users directly served by the MBS. Then, by relaxing the requirement of topology independence, we improve the proposed scheme by designing a cache placement with a subfile division dependent on network structure, which is exactly optimal in some memory size regimes. As another by-product, we also improve the existing shared-link caching scheme with shared-caches by this novel topology-dependent subfile division and placement.

研究动机与目标

• 解决缓存增强型雾无线接入网络中异构用户缓存与小基站本地缓存下的内存-负载权衡问题。
• 设计一种与拓扑无关的跨文件编码放置方案，支持有缓存与无缓存用户。
• 通过一种新型的具有共享缓存的D2D缓存方案，实现高效的跨小基站通信，以降低传输负载。
• 通过引入依赖拓扑的子文件划分与放置，改进现有共享链路缓存方案。
• 通过基于网络拓扑的结构化子文件划分，在特定内存范围内实现最优性。

提出的方法

• 提出一种新型的跨文件编码缓存放置方案，仅依赖于小基站数量与宏基站服务的用户数，而不依赖于详细网络拓扑。
• 设计一种基于共享缓存的小基站间D2D缓存方案，以实现协作传输并降低负载。
• 利用基于网络结构的子文件划分，优化编码放置方案，并在高内存场景下提升性能。
• 证明当每个小基站服务的用户数相同时，所提方案在常数因子范围内达到最优。
• 通过利用与拓扑无关的放置方案，将该方案适配至具有异构缓存大小的编码缓存系统。
• 通过引入依赖拓扑的子文件划分，增强现有基于共享缓存的共享链路缓存方案，实现更优的负载降低。

实验结果

研究问题

• RQ1 如何设计跨文件编码缓存，以在具有异构用户缓存的雾无线接入网络中最小化传输负载？
• RQ2 在小基站具备本地缓存且宏基站直接服务部分用户的情况下，缓存增强型雾无线接入网络的最优内存-负载权衡是什么？
• RQ3 在最坏情况需求下，小基站间具有共享缓存的D2D缓存方案能否实现次优性能？
• RQ4 与拓扑无关的放置相比，依赖拓扑的子文件划分在性能上有哪些提升？
• RQ5 所提出的放置方案在标准编码缓存系统中，针对异构缓存大小，其复用程度如何？

主要发现

• 所提出的跨文件编码放置方案在最坏情况内存-负载权衡下，达到常数因子范围内的次优性能。
• 该方案适用于具有异构缓存大小的编码缓存系统，包括无缓存用户。
• 与拓扑无关的放置仅依赖于小基站数量与宏基站服务的用户数，简化了在不同拓扑结构中的部署。
• 通过引入依赖拓扑的子文件划分，该方案在某些内存范围内实现了精确最优性。
• 小基站间具有共享缓存的D2D缓存方案通过协作传输降低了传输负载。
• 所提方法通过整合网络结构感知的子文件划分，改进了现有基于共享缓存的共享链路缓存方案。