The reference operator implements the basic features expected from a MyData Operator. Grasp the theory and get familiar with the internals of the reference code before starting hands-on deployment.
The deployment is easy thanks to the Docker-containerised components. Make sure you've got enough punch in your server or virtual machine so the memory doesn't run out - 2GB RAM/10GB disk is a good starting level.
Operator needs also a frontend service for its users to operate their privacy settings via their own accounts. The reference implementation is in no way complete compared to what a commercial operator would host for their users - it's a rough starting point. We have made available open source material for getting on with e.g. operator UI development.
Operating an operator means fixing multiple architecture elements and processes beyond the code, both at the end-user and back-office side of things. How do you authenticate the users? Is single-sign-on authentication available for your linked services? How will you screen services prior to accepting them to your service catalogue? Will some services appear for all MyData Account holders (pre-linking)? Are you serving only particular types of services and data (public, private, health, finance, retail, etc.)? SDK cannot yet help on all these decisions but components can be developed further in private projects.
The SDK is implemented in Python 2.7. We acknowledge that not all organisations or developers are comfortable with Python, so all interfaces are REST-based thus enabling easy integration with other proprietary components - or easy component replacement for improved performance or efficiency.