MoEless: Efficient MoE LLM Serving via Serverless Computing

TL;DR

MoEless introduces a serverless framework for efficient MoE large language model serving, reducing latency and cost by balancing expert loads and leveraging serverless computing.

Contribution

It is the first to address expert load imbalance in serverless MoE LLM serving with lightweight load predictors and optimized expert scaling strategies.

Findings

Reduces inference latency by 43%

Cuts inference cost by 84%

Improves GPU utilization and load balancing

Abstract

Large Language Models (LLMs) have become a cornerstone of AI, driving progress across diverse domains such as content creation, search and recommendation systems, and AI-assisted workflows. To alleviate extreme training costs and advancing model scales, Mixture-of-Experts (MoE) has become a popular backbone for modern LLMs, which are commonly served in distributed deployment using expert parallelism (EP). However, MoE's sparse activation mechanism leads to severe expert load imbalance, where a few experts become overloaded while others remain idle, resulting in expert stragglers that inflate inference latency and serving cost. Existing expert load balancing solutions assume static resource configurations on serverful infrastructures, limiting expert scalability and elasticity, and resulting in either costly real-time expert swapping or degraded generation quality. We present MoEless, the first serverless MoE serving framework that mitigates expert load imbalance and accelerates inference via serverless experts. MoEless employs lightweight, layer-aware predictors to accurately estimate incoming expert load distributions and proactively identify stragglers. We design optimized expert scaling and placement strategies to maximize function locality, improve GPU utilization, and balance loads across experts and GPUs. MoEless is prototyped on top of Megatron-LM and deployed on an eight-GPU testbed. Experiments with open-source MoE models and real-world workloads show that MoEless reduces inference latency by 43% and inference cost by 84% compared to state-of-the-art solutions.