通过版本与指纹追踪，实现企业文档中答案的精准复用。
FinCacheServe: Dependency-Consistent Answer Reuse for Cost-Efficient RAG Serving over Mutable Enterprise Documents
- 用文档版本、证据指纹等构建答案索引，确保依赖一致性。
- 在2230次请求中跳过53.27%的LLM调用，无过时输出。
- 适合需高效处理动态企业文档的RAG服务场景。
面向可变企业文档的检索增强生成服务中，相同语义请求重复执行。答案复用可减少GPU密集型生成负担，但缓存需保持依赖一致性。FinCacheServe将每个生成答案视为由企业意图索引的服务对象，受文档版本、证据指纹、工具指纹、模型身份和解码配置保护。基于vLLM的实现评估了基于SEC财务文档的工作负载，使用Qwen2.5模型。在2,230次请求的托管7B模型轨迹中，跳过53.27%的LLM调用，未出现依赖过时输出。在三个托管32B模型操作套件种子上，跳过53.31%的544次请求，优于版本化语义缓存（38.97%）和基于溯源的复用（22.43%）。容量、后端和SLO回放显示，其缓存管理接近理想状态，支持10万条事务元数据，且相比版本化语义缓存，依赖新鲜的2秒SLO成功案例能耗降低44.30%。
原文摘要 · Abstract (English)
Retrieval-augmented generation services over mutable enterprise documents repeatedly execute semantically equivalent analysis requests. Answer reuse can remove GPU-bound generation work, yet response caches require dependency consistency when filings, evidence chunks, and tool outputs change. FinCacheServe treats each generated answer as a serving object indexed by enterprise intent and guarded by document versions, evidence fingerprints, tool fingerprints, model identity, and decoding configuration. A vLLM implementation evaluates SEC-derived financial-document workloads with Qwen2.5 models. On a 2,230-request hosted 7B trace, FinCacheServe skips 53.27% of LLM calls with zero observed dependency-stale outputs. Across three hosted 32B operator-suite seeds, it skips 53.31% of 544 requests, compared with 38.97% for versioned semantic caching and 22.43% for grounded-style reuse. Capacity, backend, and SLO replays show oracle-bounded cache management, 100k-entry transactional metadata behavior, and 44.30% lower estimated Wh per dependency-fresh 2s-SLO success than versioned semantic caching.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。