# microsoft/garnet

Garnet is a remote cache-store from Microsoft Research that offers strong performance (throughput and latency), scalability, storage, recovery, cluster sharding, key migration, and replication features. Garnet can work with existing Redis clients.

Repository: https://github.com/microsoft/garnet
Canonical: https://ross.abutalabs.com/products/garnet
Homepage: https://microsoft.github.io/garnet/
Language: C#
License: MIT
License Family: permissive
Topics: cache, cache-storage, cluster, concurrent, hash-table, key-value, larger-than-memory, low-latency, persistent, remote, scalable, store
Last push: 2026-08-26T21:49:03+00:00

## Health v2 (maintenance only)
Score: 91/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 79
- inputs: {"age_days": 1108, "days_push": 7, "days_rel": 8, "gap_med": 5, "n_releases_24m": 94}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 11956, forks 693 (observed 2026-08-28T04:10:50.548969+00:00)

## What it is
Garnet is a high-performance remote cache-store from Microsoft Research written in C#/.NET, speaking the RESP wire protocol so unmodified Redis clients can connect to it. It provides throughput and latency-focused caching with persistence, cluster sharding, replication, key migration, and larger-than-memory tiered storage via its Tsavorite engine.

## Use cases
- drop-in Redis-compatible cache server for my app
- low-latency caching with many concurrent client connections
- persistent key-value store larger than memory
- self-hosted cache with cluster sharding and replication
- use Redis clients from C# or other languages against a .NET server
- vector similarity search with a Redis-like API

## When to choose
- you want Redis-protocol caching with better throughput/scalability on many connections
- you run .NET and want an extensible, cross-platform cache server
- you need persistence, recovery, or larger-than-memory tiered storage
- you want TLS and clustering out of the box

## When to avoid
- you need 100% parity with every Redis command or Redis modules
- you depend on the Redis ecosystem tooling and managed Redis services
- you need a tiny embedded in-process cache rather than a remote store

## Facets
- artifact type: service
- maturity: active
- function: caching, database, http-server, search-engine
- domain: databases, backend, developer-tools, self-hosted
- platform: cross-platform, windows, dotnet, self-hosted
- tags: redis-compatible, resp-protocol, key-value-store, cache-store, tsavorite, cluster-sharding, replication, vector-search, microsoft-research, linux, docker

## Member repositories
- microsoft/garnet (main) score 91

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:50.548969+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 2026-08-29T17:15:02.605727+00:00, confidence not recorded.
  - readme: https://github.com/microsoft/garnet (fetched 2026-08-28T04:10:50.548969+00:00, sha c4babd7c5b63)
  - homepage: https://microsoft.github.io/garnet/ (fetched 2026-08-29T08:12:48.497001+00:00, sha 5e771de4c11c)
- Data as of 2026-08-30T08:39:29.467469+00:00.