TokenLake: A Unified Segment-level Prefix Cache Pool for Fine-grained Elastic Long-Context LLM Serving

• URL: http://arxiv.org/abs/2508.17219v1
• Date: Sun, 24 Aug 2025 05:45:16 GMT
• Title: TokenLake: A Unified Segment-level Prefix Cache Pool for Fine-grained Elastic Long-Context LLM Serving
• Authors: Bingyang Wu, Zili Zhang, Yinmin Zhong, Guanzhe Huang, Yibo Zhu, Xuanzhe Liu, Xin Jin,
• Abstract summary: We propose a unified segment-level prefix cache pool, TokenLake.<n>It uses a cache interface to expose requests' query tensors, prefixes, and cache-aware operations.<n> TokenLake can improve throughput by up to 2.6$times$ and 2.0$times$ and boost hit rate by 2.0$times$ and 2.1$times$.
• Score: 12.80179556886128
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Prefix caching is crucial to accelerate multi-turn interactions and requests with shared prefixes. At the cluster level, existing prefix caching systems are tightly coupled with request scheduling to optimize cache efficiency and computation performance together, leading to load imbalance, data redundancy, and memory fragmentation of caching systems across instances. To address these issues, memory pooling is promising to shield the scheduler from the underlying cache management so that it can focus on the computation optimization. However, because existing prefix caching systems only transfer increasingly longer prefix caches between instances, they cannot achieve low-latency memory pooling. To address these problems, we propose a unified segment-level prefix cache pool, TokenLake. It uses a declarative cache interface to expose requests' query tensors, prefix caches, and cache-aware operations to TokenLake for efficient pooling. Powered by this abstraction, TokenLake can manage prefix cache at the segment level with a heavy-hitter-aware load balancing algorithm to achieve better cache load balance, deduplication, and defragmentation. TokenLake also transparently minimizes the communication volume of query tensors and new caches. Based on TokenLake, the scheduler can schedule requests elastically by using existing techniques without considering prefix cache management. Evaluations on real-world workloads show that TokenLake can improve throughput by up to 2.6$\times$ and 2.0$\times$ and boost hit rate by 2.0$\times$ and 2.1$\times$, compared to state-of-the-art cache-aware routing and cache-centric PD-disaggregation solutions, respectively.

Related papers

• SenCache: Accelerating Diffusion Model Inference via Sensitivity-Aware Caching [75.02865981328509]
  Caching reduces computation by reusing previously computed model outputs across timesteps.<n>We propose Sensitivity-Aware Caching (SenCache), a dynamic caching policy that adaptively selects caching timesteps on a per-sample basis.<n>SenCache achieves better visual quality than existing caching methods under similar computational budgets.
  arXiv  Detail & Related papers  (2026-02-27T17:36:09Z)
• TableCache: Primary Foreign Key Guided KV Cache Precomputation for Low Latency Text-to-SQL [26.22021341849791]
  We propose precomputing table representations as KV caches offline and querying the required ones online.<n>A key aspect of our approach is the computation of table caches while preserving primary foreign key relationships between tables.
  arXiv  Detail & Related papers  (2026-01-13T17:20:55Z)
• vCache: Verified Semantic Prompt Caching [75.87215136638828]
  This paper proposes vCache, the first verified semantic cache with user-defined error rate guarantees.<n>It employs an online learning algorithm to estimate an optimal threshold for each cached prompt, enabling reliable cache responses without additional training.<n>Our experiments show that vCache consistently meets the specified error bounds while outperforming state-of-the-art static-threshold and fine-tuned embedding baselines.
  arXiv  Detail & Related papers  (2025-02-06T04:16:20Z)
• InstCache: A Predictive Cache for LLM Serving [6.076957323090607]
  Caching techniques offer opportunities to optimize the performance of Large Language Models inference engines.<n>High variability in the content and length of instructions make it rare for identical instructions to recur within a short time window.<n>We propose InstCache, a predictive caching mechanism for LLM serving systems.
  arXiv  Detail & Related papers  (2024-11-21T03:52:41Z)
• EPIC: Efficient Position-Independent Caching for Serving Large Language Models [19.510078997414606]
  Caching improves serving performance by reusing Key-Value vectors across requests.<n>Existing context caching requires exact prefixes across requests.<n>We introduce Position-Independent Caching (PIC), which enables modular reuse of KV vectors regardless of prefixes.<n>We also introduce EPIC, a serving system incorporating our new LegoLink algorithm, which mitigates the inappropriate "attention sink" effect at every document beginning.
  arXiv  Detail & Related papers  (2024-10-20T08:42:29Z)
• Compute Or Load KV Cache? Why Not Both? [6.982874528357836]
  Cake is a novel KV cache loading system that optimally utilizes both computational and I/O resources in parallel.<n> Cake achieves on average 2.6x reduction in Time to First Token (TTFT) compared to compute-only and I/O-only methods.
  arXiv  Detail & Related papers  (2024-10-04T01:11:09Z)
• Efficient Inference of Vision Instruction-Following Models with Elastic Cache [76.44955111634545]
  We introduce Elastic Cache, a novel strategy for efficient deployment of instruction-following large vision-language models. We propose an importance-driven cache merging strategy to prune redundancy caches. For instruction encoding, we utilize the frequency to evaluate the importance of caches. Results on a range of LVLMs demonstrate that Elastic Cache not only boosts efficiency but also notably outperforms existing pruning methods in language generation.
  arXiv  Detail & Related papers  (2024-07-25T15:29:05Z)
• Training-Free Exponential Context Extension via Cascading KV Cache [49.608367376911694]
  We introduce a novel mechanism that leverages cascading sub-cache buffers to selectively retain the most relevant tokens.<n>Our method reduces prefill stage latency by a factor of 6.8 when compared to flash attention on 1M tokens.
  arXiv  Detail & Related papers  (2024-06-24T03:59:17Z)
• Get More with LESS: Synthesizing Recurrence with KV Cache Compression for Efficient LLM Inference [78.65321721142624]
  We focus on a memory bottleneck imposed by the key-value ( KV) cache. Existing KV cache methods approach this problem by pruning or evicting large swaths of relatively less important KV pairs. We propose LESS, a simple integration of a constant sized cache with eviction-based cache methods.
  arXiv  Detail & Related papers  (2024-02-14T18:54:56Z)
• LQoCo: Learning to Optimize Cache Capacity Overloading in Storage Systems [17.49411699184401]
  Cache plays an important role to maintain high and stable performance in storage systems. Existing rule-based cache management methods, coupled with engineers' manual configurations, lead to frequent cache overloading. We propose a light-weight learning-based cache bandwidth control technique, called LQoCo, which can adaptively control the cache bandwidth.
  arXiv  Detail & Related papers  (2022-03-21T13:53:22Z)
• Reinforcement Learning for Caching with Space-Time Popularity Dynamics [61.55827760294755]
  caching is envisioned to play a critical role in next-generation networks. To intelligently prefetch and store contents, a cache node should be able to learn what and when to cache. This chapter presents a versatile reinforcement learning based approach for near-optimal caching policy design.
  arXiv  Detail & Related papers  (2020-05-19T01:23:51Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.