We present a prototype transponder with fast synchronized change of parameters in the case of traffic quality degradation thanks to link impairment detection and autonegotiation capabilities. This concept has been implemented and validated in real time using a field programmable gate array (FPGA) platform, a Raspberry Pi, and a commercial transponder. We introduce an autonegotiation protocol for local decision-making between two transponders. We measure the latency added by this protocol processing cores inside the FPGA. We then integrate this transponder prototype inside a network testbed, where we compare its performance in detecting and resolving a soft failure against a centralized software-defined network control and management plane. Our solution proves to be 2 times faster and shows less variability on average than a centralized one. This work is complementary to control plane decision-making and brings further flexibility with the idea in mind to promote wider use of reconfigurations toward close-to-zero-margin optical networks.

中文翻译：

---

具有用于软件定义网络的自动协商协议的实时光转发器原型

由于链路损伤检测和自动协商功能，我们提出了一种原型转发器，在流量质量下降的情况下，可以快速同步更改参数。该概念已使用现场可编程门阵列 (FPGA) 平台、Raspberry Pi 和商业转发器实时实施和验证。我们为两个转发器之间的本地决策引入了自动协商协议。我们测量由 FPGA 内的该协议处理内核增加的延迟。然后，我们将这个转发器原型集成到网络测试台中，在那里我们将其在检测和解决软故障方面的性能与集中式软件定义的网络控制和管理平面进行比较。事实证明，我们的解决方案比集中式解决方案快 2 倍，平均可变性更小。