Abstract: Most existing content caching algorithms require accurate estimation of content popularity, which is difficult to implement in a dynamic mobile network environment. This paper proposes a cloud-to-side hybrid caching strategy for content heterogeneous 5G wireless networks, and designs and optimizes the content caching location, which can be the original content server, Cloud Unit（CUs） and Base Station（BSs）. Firstly, Lyapunov optimization method is used to solve the tight coupling problem between NP-hard cache control and CU cache and BS cache control decision. It is helpful to improve and identify the hierarchy of network architecture and the membership relationship between Cus cache and BSs cache. Cloud-centric and edge-centric caching improves content caching performance and supports high average request content data rates. Secondly, the Lyapunov optimization technique is used to realize the limited service latency of all arrival rates in the constant fraction capacity area, and then the fast reading of the cached data is realized. Simulation results show that the proposed cache strategy has significant advantages in terms of average end-to-end service delay and load reduction rate.

Key words: content heterogeneity, 5G wireless network, cloud to side, hybrid cache, service latency

摘要： 现有大多数内容缓存算法需要对内容流行度的准确估计，这在动态移动网络环境中是较难实现的。提出考虑内容异构5G无线网络云对边混合缓存策略，设计优化了内容缓存位置，其可以是原始内容服务器、云单元（CUs）和基站（BSs）。采用Lyapunov优化方法解决了NP-hard缓存控制问题与CU缓存和BS缓存控制决策之间的紧密耦合问题，有助于改善和识别网络体系结构的层次性和Cus缓存与BSs缓存之间的隶属关系，同时新的分层网络架构能够通过机会性地开发以云为中心和以边缘为中心的缓存来提高内容缓存性能，支持高平均请求的内容数据速率。采用李雅普诺夫优化技术，可实现恒定分数的容量区域的所有到达率的有限服务延迟，进而实现缓存数据的快速读取。仿真结果显示，所提缓存策略在平均端到端服务延迟和负载降低率方面具有较为显著的优势。

关键词: 内容异构, 5G无线网络, 云对边, 混合缓存, 服务延迟