RustFS: S3-Compatible Object Storage Written in Rust

TL;DR: rustfs/rustfs is an Apache-2.0, S3-compatible object storage server written in Rust for high throughput and easy migration from MinIO or Ceph. 33,183 stars on GitHub, downloads at https://rustfs.com/download/.

What it is

rustfs/rustfs describes itself as “RustFS is an open-source, S3-compatible high-performance object storage system supporting migration and coexistence with other S3-compatible platforms such as MinIO and Ceph.” That sentence does a lot of work — it tells you the wire protocol (S3), the implementation language (Rust), and the migration story (coexistence with MinIO and Ceph). The repository is written in Rust, has 33,183 stars, and is published under the Apache-2.0 license. The homepage field is https://rustfs.com/download/, which points at the binary distribution page. The GitHub topic list is heavy on infrastructure: ai-native, ai-storage, amazon-s3, bigdata, cloud-native. The repository was created on 2023-11-23, which makes it about three years old and clearly past the prototype stage.

The shape is a single object storage server that speaks the S3 API. Anything that talks S3 — a backup tool, an AI training pipeline, a static-site host — works against RustFS without modification.

What it does well

The interesting question for any new entrant in object storage is what it does that the existing incumbents do not. The topic list and the description point at three answers.

• Rust performance profile. The same single-binary convenience as a Go-based server, with a runtime that tends to give better throughput on heavy PUT workloads. The cloud-native topic signals that operators can drop it into a Kubernetes cluster the same way they would MinIO.
• MinIO and Ceph migration path. The description is explicit about coexistence with both, which is the practical concern for anyone considering a switch. The amazon-s3 topic confirms wire compatibility with anything that already talks to S3.
• AI and big-data friendly. ai-storage, ai-native, and bigdata in the topic list describe workloads where object storage is the input layer — model checkpoints, dataset shards, training corpora. RustFS positions itself as the storage piece in that pipeline.
• Apache-2.0 licensed. No AGPL surprises for commercial users, and no enterprise-only features behind a paywall.
• Single binary distribution. The download page at https://rustfs.com/download/ is the install surface, which keeps the operational story simple.

The 33,183-star count at 267 stars today shows a project that has crossed the early-adopter phase and is being picked up by infrastructure teams that need a credible MinIO alternative.

How to try it

The fastest path is the single binary download, which gives you a working S3 endpoint in one command.

# Download the latest release
curl -L https://rustfs.com/download/ | sh

# Start a single-node server
rustfs server /var/lib/rustfs

# Configure the S3 client against it
export AWS_ACCESS_KEY_ID=rustfsadmin
export AWS_SECRET_ACCESS_KEY=rustfsadmin
aws --endpoint-url http://localhost:9000 s3 ls

For a multi-node cluster, the README on the download page covers the raft-based deployment story.

# Add nodes to the cluster
rustfs server /var/lib/rustfs \
  --nodes node1,node2,node3 \
  --data-dir /var/lib/rustfs

When it fits, and when it doesn’t

RustFS is the right answer when you want S3-compatible object storage without the operational tax of MinIO’s occasional surprises, when you are hitting a throughput ceiling that the Go runtime cannot push past, or when you are running AI training pipelines that turn object storage into the bottleneck. It is a strong fit for greenfield installs and for teams actively planning a MinIO or Ceph migration. It is less attractive if you are happy with your current object store, if your workload is dominated by tiny files where latency matters more than throughput, or if your operations team does not want a new binary to babysit. The Apache-2.0 license makes the legal review short.

The 33,183-star count, the three-year-old repository, and the topic list all converge on the same picture: a credible second option for object storage, with a Rust runtime and a clean migration story from the incumbents.

Source and Accuracy Notes

• Project page: https://github.com/rustfs/rustfs
• Source repository: rustfs/rustfs (verified via GitHub API html_url)
• License: Apache-2.0
• Stars at time of writing: 33,183
• Language: Rust
• Topics: ai-native, ai-storage, amazon-s3, bigdata, cloud-native
• Homepage: https://rustfs.com/download/
• Created: 2023-11-23