# Equinox: Holistic Fair Scheduling in Serving Large Language Models

arXiv: https://arxiv.org/abs/2508.16646 · HTML: https://mlsanity.com/paper/2508.16646/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Serving systems and request scheduling, one of 12 directions in Efficient LLM inference and serving.
- Builds on 22 of its 46 references that are in the map.
- First posted on arXiv 19 Aug 2025.

## Placed in

- Efficient LLM inference and serving / [Serving systems and request scheduling](https://mlsanity.com/inference/serving-systems-and-request-scheduling/)

## Builds on (most cited first, 22 of 22)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Sequence to Sequence Learning with Neural Networks](https://mlsanity.com/paper/1409.3215/) (2014)
- [SGLang: Efficient Execution of Structured Language Model Programs](https://mlsanity.com/paper/2312.07104/) (2023)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve](https://mlsanity.com/paper/2403.02310/) (2024)
- [Are Emergent Abilities of Large Language Models a Mirage?](https://mlsanity.com/paper/2304.15004/) (2023)
- [Predictability and Surprise in Large Generative Models](https://mlsanity.com/paper/2202.07785/) (2022)
- [An Empirical Study of the Non-determinism of ChatGPT in Code Generation](https://mlsanity.com/paper/2308.02828/) (2023)
- [LoongServe: Efficiently Serving Long-Context Large Language Models with Elastic Sequence Parallelism](https://mlsanity.com/paper/2404.09526/) (2024)
- [DynamoLLM: Designing LLM Inference Clusters for Performance and Energy Efficiency](https://mlsanity.com/paper/2408.00741/) (2024)
- [Parrot: Efficient Serving of LLM-based Applications with Semantic Variable](https://mlsanity.com/paper/2405.19888/) (2024)
- [Efficient Interactive LLM Serving with Proxy Model-based Sequence Length Prediction](https://mlsanity.com/paper/2404.08509/) (2024)
- [LLM Processes: Numerical Predictive Distributions Conditioned on Natural Language](https://mlsanity.com/paper/2405.12856/) (2024)
- [Mind the Memory Gap: Unveiling GPU Bottlenecks in Large-Batch LLM Inference](https://mlsanity.com/paper/2503.08311/) (2025)
- [Enabling Efficient Batch Serving for LMaaS via Generation Length Prediction](https://mlsanity.com/paper/2406.04785/) (2024)
- [Hierarchical Roofline Performance Analysis for Deep Learning Applications](https://mlsanity.com/paper/2009.05257/) (2020)
- [Multi-Bin Batching for Increasing LLM Inference Throughput](https://mlsanity.com/paper/2412.04504/) (2024)
- [KV-Runahead: Scalable Causal LLM Inference by Parallel Key-Value Cache Generation](https://mlsanity.com/paper/2405.05329/) (2024)
- [Fair-MoE: Fairness-Oriented Mixture of Experts in Vision-Language Models](https://mlsanity.com/paper/2502.06094/) (2025)

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

- [SuperInfer: SLO-Aware Rotary Scheduling and Memory Management for LLM Inference on Superchips](https://mlsanity.com/paper/2601.20309/) (2026)
- [AdaFair-MARL: Enforcing Adaptive Fairness Constraints in Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2511.14135/) (2025)

## Abstract (arXiv)

We address the limitations of current LLM serving with a dual-counter framework separating user and operator perspectives. The User Fairness Counter measures quality of service via weighted tokens and latency; the Resource Fairness Counter measures operational efficiency through throughput and GPU utilization. Since these metrics are only available post-execution, creating a scheduling paradox, we introduce a deterministic Mixture of Prediction Experts (MoPE) framework to predict user-perceived latency, output tokens, throughput, and GPU utilization. These predictions enable calculation of a unified Holistic Fairness score that balances both counters through tunable parameters for proactive fairness-aware scheduling. We implement this in Equinox, an open-source system with other optimizations like adaptive batching, and stall-free scheduling. Evaluations on production traces (ShareGPT, LMSYS) and synthetic workloads demonstrate Equinox achieves up to 1.3× higher throughput, 60% lower time-to-first-token latency, and 13% higher fairness versus VTC while maintaining 94% GPU utilization, proving fairness under bounded discrepancy across heterogeneous platforms.