# Coordinated Scheduling for MoE LLM Serving

arXiv: https://arxiv.org/abs/2606.15177 · HTML: https://mlsanity.com/paper/2606.15177/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Efficient MoE training and inference systems, one of 12 directions in Mixture-of-experts.
- Builds on 17 of its 53 references that are in the map.
- First posted on arXiv 13 Jun 2026.

## Placed in

- Mixture-of-experts / [Efficient MoE training and inference systems](https://mlsanity.com/moe/efficient-moe-training-and-inference-systems/)
- Efficient LLM inference and serving / [Serving systems and request scheduling](https://mlsanity.com/inference/serving-systems-and-request-scheduling/)

## Builds on (most cited first, 17 of 17)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Mixtral of Experts](https://mlsanity.com/paper/2401.04088/) (2024)
- [DeepSeek-V3 Technical Report](https://mlsanity.com/paper/2412.19437/) (2024)
- [SARATHI: Efficient LLM Inference by Piggybacking Decodes with Chunked Prefills](https://mlsanity.com/paper/2308.16369/) (2023)
- [Fast Distributed Inference Serving for Large Language Models](https://mlsanity.com/paper/2305.05920/) (2023)
- [Preble: Efficient Distributed Prompt Scheduling for LLM Serving](https://mlsanity.com/paper/2407.00023/) (2024)
- [Efficient Interactive LLM Serving with Proxy Model-based Sequence Length Prediction](https://mlsanity.com/paper/2404.08509/) (2024)
- [MoETuner: Optimized Mixture of Expert Serving with Balanced Expert Placement and Token Routing](https://mlsanity.com/paper/2502.06643/) (2025)
- [Prefill-Decode Aggregation or Disaggregation? Unifying Both for Goodput-Optimized LLM Serving](https://mlsanity.com/paper/2508.01989/) (2025)
- [Efficiently Serving Large Multimodal Models Using EPD Disaggregation](https://mlsanity.com/paper/2501.05460/) (2024)
- [Injecting Adrenaline into LLM Serving: Boosting Resource Utilization and Throughput via Attention Disaggregation](https://mlsanity.com/paper/2503.20552/) (2025)
- [HydraInfer: Hybrid Disaggregated Scheduling for Multimodal Large Language Model Serving](https://mlsanity.com/paper/2505.12658/) (2025)
- [FlexPipe: Adapting Dynamic LLM Serving Through Inflight Pipeline Refactoring in Fragmented Serverless Clusters](https://mlsanity.com/paper/2510.11938/) (2025)
- [Janus: Disaggregating Attention and Experts for Scalable MoE Inference](https://mlsanity.com/paper/2512.13525/) (2025)
- [Not All Prefills Are Equal: PPD Disaggregation for Multi-turn LLM Serving](https://mlsanity.com/paper/2603.13358/) (2026)
- [SeaLLM: Service-Aware and Latency-Optimized Resource Sharing for Large Language Model Inference](https://mlsanity.com/paper/2504.15720/) (2025)
- [Efficient LLM Serving on Hybrid Real-time and Best-effort Requests](https://mlsanity.com/paper/2504.09590/) (2025)

## Built on by (most cited first, 2 of 2)

- [RoutePack: Expert Placement and Attention-Aware Data Packing for MoE Reinforcement Learning](https://mlsanity.com/paper/2608.12146/) (2026)
- [Does Execution Require Target KV Fidelity? A Mixed-Fidelity KV Runtime for LLM Serving](https://mlsanity.com/paper/2609.33536/) (2026)

## Abstract (arXiv)

Serving Mixture-of-Experts (MoE) large language models (LLMs) is challenging because dynamic request workloads interact with sparse expert routing, creating both data-parallel (DP) engine imbalance and expert-level hotspots. Existing LLM serving systems typically make these decisions in isolation: frontend schedulers route requests using coarse request counters, while backend expert balancers rely mainly on aggregate expert activation counts. This separation prevents the serving system from reacting to fine-grained engine pressure, backend MoE pressure, and source-dependent expert traffic. To address this gap, we propose Gimbal, a coordinated cross-level scheduling system for efficient MoE-based LLM serving. First, Gimbal presents a fine-grained DP-engine scheduler that uses online backend pressure signals, including key-value (KV) cache usage, remaining prefill work, queue pressure, and MoE expert pressure, to dispatch requests away from overloaded engines. Inside each engine, Gimbal further applies a lightweight prefill-aware queue ordering policy with aging to reduce head-of-line blocking without output-length prediction. Second, Gimbal extends expert load balancing with online source-DP-to-expert routing statistics and uses a heuristic guided by a mixed-integer nonlinear program (MINLP) to place experts while jointly considering expert load, source-aware communication, and migration stability. Our evaluation shows that Gimbal reduces average Time To First Token (TTFT) by 42.9% and average Time Per Output Token (TPOT) by 33.3% compared with the state-of-the-art serving system vLLM, while improving high-load request throughput by 3.0%.