EUROSYS '26: Proceedings of the 21st European Conference on Computer Systems
2026年06月05日
"Next-generation datacenters require highly efficient network load balancing to manage the growing scale of artificial intelligence (AI) training and general datacenter traffic. However, existing Ethernet-based solutions, such as Equal Cost MultiPath (ECMP) and oblivious packet spraying (OPS), struggle to maintain high network utilization due to both increasing traffic demands and the expanding scale of datacenter topologies, which also exacerbate network failures. To address these limitations, we propose REPS, a lightweight decentralized per-packet adaptive load balancing algorithm designed to optimize network utilization while ensuring rapid recovery from link failures. REPS adapts to network conditions by caching good-performing paths. In case of a network failure, REPS re-routes traffic away from it in less than 100 microseconds. REPS is designed to be deployed with next-generation out-of-order transports, such as Ultra Ethernet, and uses less than 25 bytes of per-connection state regardless of the topology size. We extensively evaluate REPS in large-scale simulations and FPGA-based NICs."
这篇文章发表于 EUROSYS ‘26,主要作者来着苏黎世联邦理工,微软,以及罗马第一大学。该文章提出了一个轻量级的去中心化逐跳自适应负载均衡算法 REPS,旨在解决现有现有以太网解决方案在日益增长的AI流量和通用数据中心流量存在的问题。另外,REPS更多的要与下一代乱序传输技术(Ultra Ethernet,超以太网)一起部署,它很大程度上依赖于该技术提供的乱序接收能力。
在训练集群规模急剧增加时,我们会面临以下问题:
这些问题在实际场景中的具体体现如下:
当前解决方案的不足:
面向大规模 AI 智算 + HPC 高性能计算的开放高性能以太网新标准,他的定位是既有 InfiniBand 的超低时延 / 无损性能,同时又保留传统以太网低成本、全生态、大规模组网优势。他和 RoCE、IB verbs 这些传输协议可以说是一类东西。
该机制的关键:
*EV: 作为hash函数的输入之一