Technical Papers Archive

Ordered Chaos: Towards Adaptive Multipath Load Balancing with Precise, Fast Loss Detection


Authors: Tommaso Bonato (ETH Zürich, Microsoft Corporation); Ahmad Ghalayini and Nadeen Gebara (Microsoft Corporation); Benoit Pit-Claudel (Massachusetts Institute of Technology (MIT)); Jiayong Li and Mikhail Khalilov (ETH Zürich); Luca Piccolboni and Waleed Reda (Microsoft Corporation); Pei-Huan Tsai (Columbia University, Microsoft Corporation); Tao Ji and Abdul Kabbani (Microsoft Corporation); and Torsten Hoefler (ETH Zürich, Microsoft Corporation)

Abstract: Load balancing in best-effort datacenter fabrics faces a trade-off. Order-preserving schemes such as equal-cost multipath (ECMP) retain loss semantics but use path diversity coarsely, while packet spraying balances load well but removes the ordering signals needed for fast, precise loss detection. We present Ordered Chaos, an end-host design that preserves enough structure for reliable loss detection without giving up fine-grained multipath routing. Reactive subflow spraying (RSS) divides each flow into ordered, dynamically routed subflows. Precise fast loss detection (PFLD) combines this ordering with probes to distinguish reordering from congestion, corruption, tail, and retransmission losses within one RTT. Across simulations, AI and HPC workloads, and a DPU prototype, RSS+PFLD improves tail flow completion time by up to 3x, eliminates up to 99% of timeout-based recoveries, and accelerates collective and training communication by up to 35%.


Presentation: file

Back to Technical Papers Archive Listing