Perplexity 自研 CobbleDB 替换 DynamoDB，查询延迟降低 5 倍
原标题：Home Made CobbleDB Replaces DynamoDB at Perplexity to Cut Query Latency 5x and Reduce Cloud Storage
AI 摘要
Perplexity 将核心搜索服务层从 Amazon DynamoDB 迁移到自研的 Rust 分布式键值存储 CobbleDB，以解决向大语言模型提供多 KB 文档批次时的高延迟与高成本问题。新架构将持久化存储与热层检索解耦，使批量读取延迟降低约 5 倍，存储费用至少降低 20%。CobbleDB 由两名系统工程师配合 AI 编码智能体在两个月内完成，公司计划将其开源。
正文节选
Perplexity has transitioned its core search serving tier away from Amazon DynamoDB to CobbleDB, an internally developed distributed key-value store written in Rust. The migration addresses severe latency and cost bottlenecks that arose from serving multi-kilobyte document batches to large language models under heavy query volumes. By decoupling durable document storage from hot-tier retrieval, the engineering team achieved a fivefold reduction in batch-read latencies while lowering overall stora