Today for AI
HOT RADAR
computeHEAT 11.5°

AI CLUSTERED EVENT · 10/8/2026

Perplexity Replaces DynamoDB with Rust-based CobbleDB for LLM Search

1 reports archived1 independent sourcesupdated 10/8/2026, 2:18:35 AM
Synthesis & Latest Updates
1 Sources Cross-Validated

Perplexity migrated its core search service from Amazon DynamoDB to CobbleDB, a custom Rust-based distributed key-value store, to address latency and cost bottlenecks in high-throughput LLM retrieval. By decoupling persistent storage from the hot data layer, the team achieved a fivefold reduction in batch read latency and cut storage costs by at least 20%. The system was built in two months by two engineers working with AI coding agents and is planned for open-source release.

LATEST/Perplexity migrated its core search service from Amazon DynamoDB to CobbleDB, a custom Rust-based distributed key-value store, to address latency and cost bottlenecks in high-throughput LLM retrieval. By decoupling persistent storage from the hot data layer, the team achieved a fivefold reduction in batch read latency and cut storage costs by at least 20%. The system was built in two months by two engineers working with AI coding agents and is planned for open-source release.

TIMELINECoverage timeline

Total 1 reports · Latest first
  1. InfoQ 中文T2·78 pts

    Perplexity Replaces DynamoDB with Rust-based CobbleDB for LLM Search

    Original: 两名工程师、两个月、4 万行 Rust：DynamoDB 太贵又慢，Perplexity 决定自己造

    • LLM retrieval payloads average ~50 KB per record, causing exponential cost increases in byte-billed cloud databases compared to traditional search metadata.
    • CobbleDB utilizes a three-system architecture: Pillar for persistent state, Lorry for batch aggregation, and CobbleDB for low-latency serving.
    • An autonomous AI coding agent cluster assisted with integration testing, build monitoring, and runbook creation, significantly accelerating the development timeline.