# bytedance/terarkdb

A RocksDB compatible KV storage engine with better performance

Repository: https://github.com/bytedance/terarkdb
Canonical: https://ross.abutalabs.com/products/terarkdb
Language: C++
License: Apache-2.0
License Family: permissive
Last push: 2026-07-13T07:10:18+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 92, release rhythm 8, longevity 100
- inputs: {"age_days": 2091, "days_push": 51, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2153, forks 213 (observed 2026-08-28T04:06:19.446339+00:00)

## What it is
TerarkDB is a RocksDB-compatible embedded key-value storage engine forked from RocksDB v5.18.3, offering improved tail latency, throughput, and compression. It can be used as a drop-in replacement for RocksDB in C/C++ applications on Linux.

## Use cases
- replace rocksdb with a faster kv store
- reduce tail latency in embedded key-value storage
- migrate existing rocksdb data to a better engine
- improve compression for large-value workloads
- embed a high-performance storage engine in a c++ application

## When to choose
- you already use RocksDB on Linux and want better performance without changing your API
- your workload is read-heavy or write-heavy with large values and needs lower tail latency
- you want better compression in an embedded KV store

## When to avoid
- you need production support on platforms other than Linux
- you need language bindings beyond C/C++
- you may need to migrate data back to RocksDB, which is not supported

## Facets
- artifact type: library
- maturity: maintenance
- function: database, caching, compression
- domain: databases, developer-tools, performance
- platform: cpp, c
- tags: kv-storage, rocksdb-fork, lsmtree, storage-engine, embedded-database, linux

## Member repositories
- bytedance/terarkdb (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:19.446339+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:50:38.044929+00:00, confidence not recorded.
  - readme: https://github.com/bytedance/terarkdb (fetched 2026-08-28T04:06:19.446339+00:00, sha d998f195f43a)
- Data as of 2026-08-30T08:39:29.467469+00:00.