To meet the challenges posed by the explosive growth of mobile traffic, data caching at the network edge has been considered a key technology in future mobile networks, while the potential of device-to-device (D2D) communications in areas such as traffic offloading is also of great interest. Existing work does not have a network switching mechanism, which would ensure load balancing and improve quality of service are also ignored. Existing simulators perform poorly in terms of algorithmic compatibility, and require a high level of coding ability which are difficult to get started. In this paper, an edge simulator called SimEdgeIntel is presented for resource management that opens up detailed configuration options, enabling researchers quickly deploy mobile with edge intelligence. It supports researchers to customize the development of mobility models, caching algorithms and switching strategies. The interface-oriented system architecture helps researchers achieve cross-platform and cross-language algorithm import with machine learning techniques. In the experimental section, we perform a comprehensive evaluation of SimEdgeIntel based on real-world tracing, proving its scalability and effectiveness in terms of cache hit rate, delivery latency, and backhaul traffic, and evaluating its performance in terms of CPU and memory, respectively.

为了应对移动流量爆炸性增长带来的挑战，网络边缘的数据缓存已被认为是未来移动网络中的一项关键技术，而在流量分流等领域，设备到设备（D2D）通信的潜力是巨大的。也很感兴趣。现有工作没有网络交换机制，该机制将确保负载平衡和提高服务质量也被忽略。现有的模拟器在算法兼容性方面表现不佳，并且需要难以入门的高水平编码能力。在本文中，提出了一种名为SimEdgeIntel的边缘模拟器，用于资源管理，它打开了详细的配置选项，使研究人员可以利用边缘智能快速部署移动设备。它支持研究人员自定义移动性模型，缓存算法和切换策略的开发。面向接口的系统架构可帮助研究人员通过机器学习技术实现跨平台和跨语言算法的导入。在实验部分，我们将基于现实世界的跟踪对SimEdgeIntel进行全面评估，以证明其在高速缓存命中率，传递延迟和回程流量方面的可扩展性和有效性，并分别在CPU和内存方面评估其性能。 。