RustFS: High-performance distributed object storage for MinIO alternative (github.com)

RustFS is a new high-performance, distributed object storage project written in Rust positioning itself as an alternative to MinIO. It touts full S3 compatibility, an Apache 2.0 business-friendly license (vs. AGPL-style restrictions), and optimizations for data lake, AI and big-data workloads. Built with Rust for memory safety and low-level performance, RustFS emphasizes a simple operator experience, built-in observability (Grafana, Prometheus, Jaeger), and community-driven development with multi-architecture build support for amd64 and arm64. Key technical highlights: RustFS implements a scalable, fault-tolerant distributed architecture with S3 API compatibility so existing ML/data pipelines can plug in with little change. The project provides Docker images, a docker-compose profile that launches observability stacks, and scripts for multi-arch image builds and pushes; a web console listens on port 9000 (default admin credentials advertised). A public stress-test config referenced modest hardware (2-core Xeon, 4 GB RAM, 15 Gbps network, four 40 GB drives), and the repo claims higher throughput and safety compared with Go/C implementations thanks to Rust’s memory model. Important caveat: RustFS is rapidly evolving and explicitly flagged alpha — don’t run it in production yet — but it’s worth watching for teams seeking an open-source, Rust-native object store for AI/ML infrastructure.