# Prefill-Decode Aggregation or Disaggregation? Unifying Both for Goodput-Optimized LLM Serving

arXiv: https://arxiv.org/abs/2508.01989 · HTML: https://mlsanity.com/paper/2508.01989/

## Key facts

- Cited by 23 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 16 of its 34 references that are in the map.
- First posted on arXiv 4 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, 16 of 16)

- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Code Llama: Open Foundation Models for Code](https://mlsanity.com/paper/2308.12950/) (2023)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [Splitwise: Efficient generative LLM inference using phase splitting](https://mlsanity.com/paper/2311.18677/) (2023)
- [Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve](https://mlsanity.com/paper/2403.02310/) (2024)
- [Mooncake: A KVCache-centric Disaggregated Architecture for LLM Serving](https://mlsanity.com/paper/2407.00079/) (2024)
- [NanoFlow: Towards Optimal Large Language Model Serving Throughput](https://mlsanity.com/paper/2408.12757/) (2024)
- [Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline](https://mlsanity.com/paper/2305.13144/) (2023)
- [S3: Increasing GPU Utilization during Generative Inference for Higher Throughput](https://mlsanity.com/paper/2306.06000/) (2023)
- [Efficient Interactive LLM Serving with Proxy Model-based Sequence Length Prediction](https://mlsanity.com/paper/2404.08509/) (2024)
- [InferCept: Efficient Intercept Support for Augmented Large Language Model Inference](https://mlsanity.com/paper/2402.01869/) (2024)
- [JITServe: SLO-aware LLM Serving with Imprecise Request Information](https://mlsanity.com/paper/2504.20068/) (2025)
- [Injecting Adrenaline into LLM Serving: Boosting Resource Utilization and Throughput via Attention Disaggregation](https://mlsanity.com/paper/2503.20552/) (2025)
- [Revisiting Service Level Objectives and System Level Metrics in Large Language Model Serving](https://mlsanity.com/paper/2410.14257/) (2024)

## Built on by (most cited first, 23 of 23)

- [PDTrim: Targeted Pruning for Prefill-Decode Disaggregation in Inference](https://mlsanity.com/paper/2509.04467/) (2025)
- [FairBatching: Fairness-Aware Batch Formation for LLM Inference](https://mlsanity.com/paper/2510.14392/) (2025)
- [TokenLake: A Unified Segment-level Prefix Cache Pool for Fine-grained Elastic Long-Context LLM Serving](https://mlsanity.com/paper/2508.17219/) (2025)
- [From Tokens to Layers: Redefining Stall-Free Scheduling for MoE Serving with Layered Prefill](https://mlsanity.com/paper/2510.08055/) (2025)
- [HFX: Joint Design of Algorithms and Systems for Multi-SLO Serving and Fast Scaling](https://mlsanity.com/paper/2508.15919/) (2025)
- [Astrolabe: Balancing Load in LLM Serving with Randomized Prediction-Guided Scheduling](https://mlsanity.com/paper/2508.03611/) (2025)
- [RelayGR: Scaling Long-Sequence Generative Recommendation via Cross-Stage Relay-Race Inference](https://mlsanity.com/paper/2601.01712/) (2026)
- [Revisiting Disaggregated Large Language Model Serving for Performance and Energy Implications](https://mlsanity.com/paper/2601.08833/) (2025)
- [Observation, Not Prediction: Conversation-Level Disaggregated Scheduling for Agentic Serving](https://mlsanity.com/paper/2606.01839/) (2026)
- [Coordinated Scheduling for MoE LLM Serving](https://mlsanity.com/paper/2606.15177/) (2026)
- [Towards Load-Aware Prefill Deflection for Disaggregated LLM Serving](https://mlsanity.com/paper/2607.02043/) (2026)
- [SMetric: Rethink LLM Scheduling for Serving Agents with Balanced Session-centric Scheduling](https://mlsanity.com/paper/2607.08565/) (2026)
- [The Price of Anarchy in Disaggregated Inference](https://mlsanity.com/paper/2606.17081/) (2026)
- [ExpertPlex: A High-Goodput Disaggregated Serving System for MoE LLMs with Adaptive Persistent Kernels](https://mlsanity.com/paper/2607.18002/) (2026)
- [Crossflow: Prefill-Decode Elasticity for Agentic LLM Serving](https://mlsanity.com/paper/2609.27085/) (2026)
- [ProServe: Unified Multi-Priority Request Scheduling for LLM Serving](https://mlsanity.com/paper/2512.12928/) (2025)
- [Balancing Sustainability And Performance: The Role Of Small-Scale LLMs In Agentic Artificial Intelligence Systems](https://mlsanity.com/paper/2601.19311/) (2026)
- [LLM Inference on IMC-NoC Architecture with Balanced Dataflow and Fine-Grained Parallelism](https://mlsanity.com/paper/2609.00857/) (2026)
- [Toward Sustainable Distributed LLM Inference: A Systems Synthesis and Research Agenda for an Energy-, Carbon-, and Cache-Aware llm-d Control Plane](https://mlsanity.com/paper/2609.05565/) (2026)
- [PipeSwift: Revisiting Pipeline Parallelism for Large-Scale Completion-Oriented Agentic LLM Serving](https://mlsanity.com/paper/2609.16491/) (2026)
- [FluidPD: In-Place Elasticity for SLO-Aware Prefill-Decode Disaggregated LLM Serving](https://mlsanity.com/paper/2610.06917/) (2026)
- [Prorsa: Unleashing the Potential of Prefill-Decode Disaggregation in Agentic Reinforcement Learning](https://mlsanity.com/paper/2610.16592/) (2026)
- [Scaling LLM Inference via Parallelism: Throughput-Latency Trade-off](https://mlsanity.com/paper/2610.19405/) (2026)

## Abstract (arXiv)

An ongoing debate considers whether prefill-decode (PD) aggregation or disaggregation is superior for serving large language models (LLMs). This has driven optimizations for both approaches, each showing distinct advantages. This paper compares PD aggregation and disaggregation, showing that each excels under different service-level objectives (SLOs): aggregation is optimal for tight time-to-first-token (TTFT) and relaxed time-per-output-token (TPOT), while disaggregation excels for strict TPOT and relaxed TTFT. However, under balanced TTFT and TPOT SLOs, neither approach delivers optimal goodput. This paper proposes TaiChi, an LLM serving system that unifies PD disaggregation and aggregation for optimal goodput under any combination of TTFT and TPOT SLOs. TaiChi uses a unified disaggregation-aggregation architecture with differentiated-capability GPU instances: prefill-heavy (fast prefill, high-interference decode) and decode-heavy (low-interference decode, slow prefill). Three configurable sliders control the ratio between these instances and their chunk sizes. TaiChi adapts to various SLO regimes by adjusting sliders. When TTFT constraints are tight, TaiChi resembles a PD aggregation configuration; when TPOT dominates, it adapts toward PD disaggregation. Crucially, under balanced SLOs, TaiChi enables a hybrid mode for superior goodput. The key innovation behind this hybrid mode is latency shifting: selectively reallocating GPU resources from requests that meet SLOs to those at risk of violation, maximizing the number of SLO-satisfied requests. This fine-grained latency shifting is orchestrated by two scheduling mechanisms: flowing decode scheduling to control TPOTs and length-aware prefill scheduling to manage TTFTs, which jointly optimize request assignment. Our experiments show TaiChi improves goodput by up to 77% over state-of-the-art systems under balanced TTFT and TPOT SLOs.