First Field-Trial Demonstration of L4 Autonomous Optical Network for Distributed AI Training Communication: An LLM-Powered Multi-AI-Agent Solution

Abstract: We demonstrate the first cross - domain cross - layer level - 4 autonomous optical network via a multi - AI - agent system. Field trials show ~ 9 8 % task completion rate across the distributed AI training lifecycle -- 3.2 higher than single agents using state - of - the - art LLMs. Since collaborative resource utilization across distributed facilities is essential for training workloads, t his evolution introduces significant complexity in network management, as controller s must operate across multiple domains, spanning from intra - and inter - datacenter s to long - haul wide area networks . Moreover, distributed training impose s stringent reliability requirements as it should restart from the checkpoint if a failure happens [ 2 ] . T herefore, in terms of distributed training communications, resilient operations and rapid fault recovery are essential .