SlateDB

Store and query key-value data directly on cloud object storage without local disks

Pricing model

Free

Rating

★ Mixed Reviews

Time to value

Quick Setup (< 1 hour)

You can use SlateDB to store and retrieve key-value data directly against cloud object storage systems like AWS S3, Google Cloud Storage, or Cloudflare R2. It operates as an embedded Log-Structured Merge-tree engine that eliminates the need for expensive local disk storage. You can deploy it as a single-writer system with multiple readers for scaling read workloads. It offers snapshot isolation, transactions, zero-copy cloning, and customizable caching layers to balance write latency against API request expenses across various programming languages.

What SlateDB does

Put and get key-value pairs directly from object storageCreate zero-copy clones and database branches for testingConfigure local memory and disk caches to speed up readsAdjust write flush intervals to balance cost and durabilitySeparate read replicas from primary write instancesRun compaction processes as isolated external servicesImplement optimistic concurrency checks on concurrent database transactionsZero-disk architecture that writes and reads data directly on object storesFencing protocol for safe single-writer and multi-reader architecturesZero-copy database cloning, branching, and rescaling operationsTunable latency and API request cost trade-offs for write pathsHybrid memory and local disk caching to avoid cold read latencyTransactional semantics with snapshots and optimistic concurrency controlsNative language bindings for Rust, Go, Java, Node.js, and Python