Skip to main content
Mistrive is a distributed storage platform for storing and managing large artifacts at scale. The self-hosted deployment consists of three core components that work together to deliver a complete artifact management solution.

Components

How it works

DFS forms the storage layer.
  • It splits files into 1 MiB chunks and distributes them across storage nodes with replication or Reed-Solomon erasure coding.
  • The Metadata Service tracks file locations, directory structures, and stripe placement in FoundationDB.
Artifact Store sits on top of DFS.
  • When you docker push an image, it receives the layers and manifests via the Docker Registry API, then persists them through DFS.
  • A scan worker examines pushed artifacts for vulnerabilities.
Platform ties everything together with authentication, organization hierarchies, project-level access control, and service accounts. It exposes gRPC and REST APIs alongside a management UI.

Deployment requirements

All components require access to a FoundationDB cluster for metadata storage. DFS additionally needs local storage on each node for chunk data. A minimal deployment runs all services on a single machine. Larger deployments distribute services across multiple nodes for availability.
FoundationDB provides the transactional foundation that enables atomic operations across the distributed system. You’ll need a running FoundationDB cluster before deploying any Mistrive component.

Get started

The quickstart guide walks through deploying all three components in a development environment. You’ll have a working self-hosted Mistrive instance running locally within minutes.

Mistrive Console in light mode

Quickstart

Deploy DFS, Artifact Store, and Platform in your local environment. Start pushing container images in under 10 minutes.