Supplementary information for Combining contract theory and Lyapunov optimization for content sharing with edge caching and device-to-device communications

Article abstractThe paper proposes a novel framework based on the contract theory and Lyapunov optimization for content sharing in a wireless content delivery network (CDN) with edge caching and device-to-device (D2D) communications. The network is partitioned into a set of clusters. In a cluster, users can share contents via D2D links in coordination with the cluster head. Upon receiving the content request from any user in its cluster, the cluster head either delivers the content itself or forwards the request to another node, i.e., a base station (BS) or another user in the cluster. The content access at the BS and in each cluster is modeled as a queuing system, where arrivals represent the content requests directed to respective nodes. The objective is to assign content delivery nodes to stabilize all queues while minimizing the time-averaged network cost given incomplete information about content sharing costs of the users and unknown distribution of the network state defined by users' locations and their cached/requested content. The proposed framework allows the users to truthfully reveal their content sharing expenditures, minimize the time-averaged network cost and stabilize the queuing system representing the CDN. Based on this framework, a distributed content access and delivery algorithm where the node assignments are made by every cluster head independently is developed. It is shown that the algorithm converges to the optimal policy with the trade-off in total queue backlog and achieves a superior performance compared with some other D2D content sharing policies.© 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

文章摘要：本文提出一种基于契约理论（contract theory）与李雅普诺夫优化（Lyapunov optimization）的新型框架，用于支持具备边缘缓存与设备到设备（D2D）通信能力的无线内容分发网络（CDN, Content Delivery Network）中的内容共享。
该网络被划分为多个簇。在单个簇内，用户可通过与簇头协作，经由D2D链路共享内容。当簇头接收到簇内任意用户的内容请求时，其可自行传输该内容，或将请求转发至其他节点，即基站（BS, base station）或簇内其他用户。
基站与各簇内的内容接入过程均被建模为排队系统，其中到达流对应指向各节点的内容请求。本文的优化目标为：在用户内容共享成本信息不完全、且由用户位置、缓存内容与请求内容构成的网络状态分布未知的条件下，分配内容传输节点以稳定所有队列，同时最小化时间平均网络成本。
所提框架可使用户如实披露其内容共享成本，实现时间平均网络成本的最小化，并稳定代表CDN的排队系统。基于该框架，本文开发了一种分布式内容接入与传输算法，该算法由各簇头独立完成节点分配任务。
研究表明，该算法可收敛至在总队列积压与时间平均网络成本之间取得最优权衡的最优策略，且相较于其他若干D2D内容共享策略，其性能更优。
© 2020 IEEE。允许个人使用本材料。若需以任何当前或未来媒介进行其他使用，包括为广告或促销目的重印/再发布本材料、创建新的汇编作品、转售或将材料分发至服务器或列表，或在其他作品中复用本作品的任何受版权保护的组件，均需获得IEEE的许可。