Brito Jorge Andrés, Moreno José Ignacio, Contreras Luis M
Departamento de Ingeniería de Sistemas Telemáticos, ETSI de Telecomunicación, Universidad Politécnica de Madrid, 28040 Madrid, Spain.
Telefónica Innovación Digital, 28010 Madrid, Spain.
Sensors (Basel). 2025 Apr 9;25(8):2375. doi: 10.3390/s25082375.
Next-generation networks, particularly 5G and beyond, face rising energy demands that pose both economic and environmental challenges. In this work, we present a traffic management scheme leveraging programmable data planes and an SDN controller to achieve energy proportionality, matching network resource usage to fluctuating traffic loads. This approach integrates flow monitoring on programmable switches with a dynamic power manager in the controller, which selectively powers off inactive switches. We evaluate this scheme in an emulated edge environment across multiple urban traffic profiles. Our results show that disabling switches not handling traffic can significantly reduce energy consumption, even under relatively subtle load variations, while maintaining normal network operations and minimizing overhead on the control plane. We further include a projected savings analysis illustrating the potential benefits if the solution is deployed on hardware devices such as Tofino-based switches. Overall, these findings highlight how data plane-centric, energy-aware traffic management can make 5G-and-beyond edge infrastructures both sustainable and adaptable for future networking needs.
下一代网络，尤其是5G及以后的网络，面临着不断增长的能源需求，这带来了经济和环境方面的挑战。在这项工作中，我们提出了一种流量管理方案，该方案利用可编程数据平面和软件定义网络（SDN）控制器来实现能源比例性，使网络资源使用与波动的流量负载相匹配。这种方法将可编程交换机上的流量监控与控制器中的动态电源管理器集成在一起，该管理器会选择性地关闭不活动的交换机。我们在跨多个城市交通概况的模拟边缘环境中评估了该方案。我们的结果表明，即使在相对细微的负载变化情况下，禁用不处理流量的交换机也能显著降低能耗，同时保持正常的网络运行并将控制平面上的开销降至最低。我们还进行了预计节省分析，说明了如果将该解决方案部署在基于Tofino的交换机等硬件设备上可能带来的好处。总体而言，这些发现凸显了以数据平面为中心、具备能源意识的流量管理如何能使5G及以后的边缘基础设施既可持续又能适应未来的网络需求。
