NeFut Logo NeFut
Admin Login

[CS.AI] Towards Efficient Large Language Model Serving: System-Aware KV Cache Optimization

Published at: 2026-07-11 22:00 Last updated: 2026-07-13 08:40
#algorithm #Machine Learning #Open Source

Abstract

Despite the rapid advancements of large language models (LLMs), LLM serving systems remain memory-intensive and costly. The key-value (KV) cache, which stores KV tensors during autoregressive decoding, is crucial for enabling low-latency, high-throughput LLM inference serving. This survey focuses on system-aware KV infrastructure for serving LLMs (abbreviated as sKis).

We revisit recent work from a system behavior perspective, organizing existing efforts into three dimensions: execution and scheduling (temporal), placement and migration (spatial), and representation and retention (structural). Furthermore, we analyze cross-behavior co-design affinity and behavior-objective links, highlighting future opportunities.

Our work systematizes a rapidly evolving area, providing a foundation for understanding and innovating KV cache designs in modern LLM serving infrastructure.

Blogger's Review: This paper offers a fresh perspective on KV cache analysis from a system behavior standpoint, facilitating the development of efficient serving for large language models. It points out future optimization directions and encourages deeper discussions on practical implementations of these theories.

Original Source: https://arxiv.org/abs/2607.08057

[h] Back to Home