# Delay-Adaptive Speculation Control for Low-Latency Edge-Cloud LLM Inference

arXiv: https://arxiv.org/abs/2606.20591 · HTML: https://mlsanity.com/paper/2606.20591/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Early exit and edge inference, one of 12 directions in Efficient LLM inference and serving.
- Builds on 20 of its 29 references that are in the map.
- First posted on arXiv 17 May 2026.

## Placed in

- Efficient LLM inference and serving / [Early exit and edge inference](https://mlsanity.com/inference/early-exit-and-edge-inference/)

## Builds on (most cited first, 20 of 20)

- [Fast Inference from Transformers via Speculative Decoding](https://mlsanity.com/paper/2211.17192/) (2022)
- [Accelerating Large Language Model Decoding with Speculative Sampling](https://mlsanity.com/paper/2302.01318/) (2023)
- [Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads](https://mlsanity.com/paper/2401.10774/) (2024)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty](https://mlsanity.com/paper/2401.15077/) (2024)
- [Splitwise: Efficient generative LLM inference using phase splitting](https://mlsanity.com/paper/2311.18677/) (2023)
- [SpecInfer: Accelerating Generative Large Language Model Serving with Tree-based Speculative Inference and Verification](https://mlsanity.com/paper/2305.09781/) (2023)
- [Blockwise Parallel Decoding for Deep Autoregressive Models](https://mlsanity.com/paper/1811.03115/) (2018)
- [REST: Retrieval-Based Speculative Decoding](https://mlsanity.com/paper/2311.08252/) (2023)
- [Mobile Edge Intelligence for Large Language Models: A Contemporary Survey](https://mlsanity.com/paper/2407.18921/) (2024)
- [Distributed Deep Neural Networks over the Cloud, the Edge and End Devices](https://mlsanity.com/paper/1709.01921/) (2017)
- [Online Speculative Decoding](https://mlsanity.com/paper/2310.07177/) (2023)
- [JointDNN: An Efficient Training and Inference Engine for Intelligent Mobile Cloud Computing Services](https://mlsanity.com/paper/1801.08618/) (2018)
- [SpecDec++: Boosting Speculative Decoding via Adaptive Candidate Lengths](https://mlsanity.com/paper/2405.19715/) (2024)
- [TETRIS: Optimal Draft Token Selection for Batch Speculative Decoding](https://mlsanity.com/paper/2502.15197/) (2025)
- [SLED: A Speculative LLM Decoding Framework for Efficient Edge Serving](https://mlsanity.com/paper/2506.09397/) (2025)
- [Fast and Cost-effective Speculative Edge-Cloud Decoding with Early Exits](https://mlsanity.com/paper/2505.21594/) (2025)
- [Correctness Forensics for Batch Speculative Decoding: Diagnosing the Ragged Tensor Problem](https://mlsanity.com/paper/2510.22876/) (2025)
- [FlexSpec: Frozen Drafts Meet Evolving Targets in Edge-Cloud Collaborative LLM Speculative Decoding](https://mlsanity.com/paper/2601.00644/) (2026)
- [ConfigSpec: Profiling-Based Configuration Selection for Distributed Edge--Cloud Speculative LLM Serving](https://mlsanity.com/paper/2604.09722/) (2026)

## Built on by (most cited first, 1 of 1)

- [AdaSpark: Adaptive DSpark with Online Learning for Tree Verification and N-gram Fill](https://mlsanity.com/paper/2610.05774/) (2026)

## Abstract (arXiv)

Speculative decoding accelerates large language model (LLM) inference by using a lightweight draft model to propose tokens and a larger target model to verify them in parallel. In distributed edge-cloud inference, however, draft length must be controlled online: longer drafts amortize communication delay but reduce token acceptance, whereas shorter drafts preserve acceptance but trigger more communication rounds. We formulate this tradeoff as a ratio-type optimal stopping problem and prove that the optimal draft length is a finite delay-monotone threshold. The analysis identifies a critical delay below which single-token speculation is optimal and shows that the optimal length grows only logarithmically with communication delay. For time-varying networks, we extend the model to Markov-modulated channels and establish, under a bounded horizon and monotone stopping-region conditions, a state-dependent threshold policy. For unknown environments, we propose UCB-SpecStop, an online control algorithm with gap-free and gap-dependent expected regret bounds of O(Lmax√KmaxTlog(KmaxT)) and O(sumk:Deltak>0Lmax2log(KmaxT)/Deltak). We implement the method on a real edge-cloud testbed with a Jetson Orin Nano Super edge node and an RTX~3090 Ti cloud node, using Qwen and Llama draft--target pairs. Experiments validate the predicted phase transition, with transition points near 83~ms and 111~ms. Qwen matches the geometric prediction, while Llama requires empirical-prefix calibration due to heavy-head acceptance. Across the tested delay grid, UCB-SpecStop reduces per-token latency over SpecDec++ by up to 22.4%, approaches an offline oracle within 0.2--2.4% in communication-dominated regimes, improves over naive UCB by up to 7.5%, removes the 14.0--18.7% gap caused by static tuning under delay drift, and gains 3.0--6.8% with contextual channel-state information.