Device-to-device (D2D) transmissions in wireless edge networks are promising for optimizing system-wide energy consumption and improving system service capacity. Cooperative content caching similarly serves efficiency goals for data-intensive applications in edge networks. In this work, we propose two cooperative cache replacement algorithms in D2D networks to support these techniques: i) distance-based ii) priority-class based. Video content dissemination in an edge network is our main use-case. In such content traffic, video characteristics have a significant impact on the system behavior. Therefore, we also investigate the effect of content scene change dynamics in our system. Distance based cooperation outperforms $LRU$ , $MIN$ - $ACC$ and $SXO$ in terms of goodput while priority-class based approach consumes less energy than $MIN$ - $ACC$ with almost the same consumption as $LRU$ , especially under fast-changing scene regime. Besides, it is energy-wise slightly more rewarding than $SXO$ for the fastest-changing scene case.

中文翻译：

---

D2D边缘网络中的协同缓存和视频特性

无线边缘网络中的设备到设备 (D2D) 传输有望优化系统范围的能耗和提高系统服务能力。协作内容缓存类似地为边缘网络中的数据密集型应用程序提供了效率目标。在这项工作中，我们在 D2D 网络中提出了两种协作缓存替换算法来支持这些技术：一世） 基于距离的 ii)基于优先级。边缘网络中的视频内容传播是我们的主要用例。在这样的内容流量中，视频特性对系统行为有显着影响。因此，我们还研究了我们系统中内容场景变化动态的影响。基于距离的合作表现优于 $LRU$ , $MIN$ —— $ACC$ 和 $SXO$ 在产量方面，而基于优先级的方法消耗的能量少于 $MIN$ —— $ACC$ 消耗量几乎相同 $LRU$ ，尤其是在快速变化的场景制度下。此外，它在能源方面比 $SXO$ 对于变化最快的场景案例。