Development Roadmap

We are working hard to release all the major capabilities of the platform in the open source edition of TRAC. This list is an overview of the major capabilities that have been released so far, as well as those that are in progress and planned for the near future.

You can get up to date information from our high level milestones.

Core features

• 0.1: Metadata Service ✓
• 0.2: Model Runtime for Python ✓
• 0.3: Data Service ✓
• 0.4: Local execution, TRAC D.A.P. moves to FINOS ✓
• 0.5: Orchestrator, gateway & authentication ✓
• 0.6: Cloud storage & data enhancements ✓
• 0.7: Policy service, fine-grained access control
• 0.8: Advanced analytics

You can see details of the these components and the capabilities they provide on the platform components page.

Post 1.0 features

• TRAC Shell - interactive model development tools with remote execution
• Event-based models - spot decisions or classifications based on individual data points

Plugins

These plugins are available today:

• Metadata (SQL / cloud SQL): H2 (local), MySQL, MariaDB, PostgreSQL, Microsoft SQL Server
• Storage: Local, remote (mounted filesystem), AWS S3 (includes NetApp), Google GCS, Azure BLOB
• Execution: Local, remote (SSH)
• Authentication: Guest, basic / built-in

These plugins are planned as part of the core platform:

• Storage: HDFS
• Execution: Kubernetes, AWS Batch, Hadoop YARN
• Authentication: SAML, OpenID

You can also write your own plugins for any of the capabilities exposed by the TRAC platform. If you would like a plugin and think it would also be useful to others please get in touch and we can discuss adding it to the core platform.

Other assets

As well as these functional areas, we are constantly working on deployment patterns, developer experience, documentation, compliance and enterprise readiness. Our goal is to make the platform as easy to use and deploy as possible, and answer most enterprise compliance requirements out of the box.