论文标题
通过随机缓存的多层异质网络中多质量视频的服务延迟分析和优化
Analysis and Optimization of Service Delay for Multi-quality Videos in Multi-tier Heterogeneous Network with Random Caching
论文作者
论文摘要
为了最大程度地减少服务延迟，我们建议在设备到设备（D2D）辅助的异质网络中提出一种新的随机缓存方案。为了支持多媒体视频服务的多元化查看质量，每个视频文件都通过可扩展的视频编码（SVC）编码为基础层（BL）和多个增强层（ELS）。包括BL和几个EL在内的超级层传输给每个用户。当用户由D2D助手，小型细胞基站（SBS）和宏观细胞基站（MBS）服务时，我们通过得出成功的传输概率来定义和量化多质量视频的服务延迟。我们制定了一个延迟最小化问题，但要受D2D帮助者和SBS的缓存大小有限。揭示了问题的最佳解决方案的结构，然后设计了改进的标准梯度投影方法，以有效地获得解决方案。理论分析和蒙特卡洛模拟都验证了成功的传输概率。与三个基准缓存策略相比，拟议的基于SVC的随机缓存方案在减少服务延迟方面表现出色。
Aiming to minimize service delay, we propose a new random caching scheme in device-to-device (D2D)-assisted heterogeneous network. To support diversified viewing qualities of multimedia video services, each video file is encoded into a base layer (BL) and multiple enhancement layers (ELs) by scalable video coding (SVC). A super layer, including the BL and several ELs, is transmitted to every user. We define and quantify the service delay of multi-quality videos by deriving successful transmission probabilities when a user is served by a D2D helper, a small-cell base station (SBS) and a macro-cell base station (MBS). We formulate a delay minimization problem subject to the limited cache sizes of D2D helpers and SBSs. The structure of the optimal solutions to the problem is revealed, and then an improved standard gradient projection method is designed to effectively obtain the solutions. Both theoretical analysis and Monte-Carlo simulations validate the successful transmission probabilities. Compared with three benchmark caching policies, the proposed SVC-based random caching scheme is superior in terms of reducing the service delay.
