Broadcast Rate Limits in Wi-Fi: A Forgotten Bottleneck for Collaborative Edge LLM Inference
2026-08-03 • Networking and Internet Architecture
Networking and Internet Architecture
AI summaryⓘ
The authors explore ways to run large language models on small edge devices by using a technique called Mixture-of-Experts (MoE), which spreads work across many nodes. They suggest using UDP broadcast instead of traditional methods to send data more efficiently between these devices, achieving about 1.4 times faster performance in wired setups. However, they found that current Wi-Fi limits broadcast speeds at 54 Mbps, much lower than needed for this use, especially at different distances. Through simulations, the authors recommend updating Wi-Fi standards to support higher broadcast speeds, treating broadcasts as important for data transfer, not just control signals.
Large Language ModelsMixture-of-ExpertsEdge ComputingUDP BroadcastNCCLTCPIEEE 802.11Wi-Fi BroadcastNS-3 SimulationDistributed Inference
Authors
Liujianfu Wang, Yuyang Du, Shiqi Xu, Soung Chang Liew
Abstract
LLM deployment is migrating from data centers to edge devices, where Mixture-of-Experts (MoE) models offer a promising path: sparse expert activation allows the model to be spread across multiple low-cost edge nodes. Distributed MoE inference repeatedly dispatches embeddings from one main node to many workers - a one-to-many pattern poorly served by the sequential unicasts of mainstream stacks (NCCL, TCP), yet naturally matched by UDP broadcast. We propose a UDP broadcast method for collaborative edge MoE inference, augmented with timeout-driven retransmission exploiting near deterministic latency in distributed MoE for reliability and unordered result gathering for robustness to expert mispredictions, yielding a consistent 1.4x speedup over NCCL and TCP on a wired 8-node cluster. In wireless settings, however, we uncover a deeper, long-forgotten bottleneck: IEEE 802.11 caps broadcast rates at 54 Mbps regardless of physical-layer capacity - a legacy policy built for sparse control traffic, not edge AI. NS-3 simulations at distances 1m, 2m and 5m show that the optimal rates are much higher (64x, 43x, and 32x, respectively) than the 54 Mbps cap applied in standard. Thus, we argue that broadcast is no longer a control-plane relic: it is time for Wi-Fi standards to treat it as a high-throughput data-plane citizen.