The decision has been to deliver a lot of the sync functionality via a framework approach. There has been no technical discussion on how this will actually look.

The high level purpose of this task is to

• evaluate all of the existing components/functionality inside the data sync server and the data-sync-gql core repos
• Experiment how we can extract those pieces into a framework / set of individual libraries

As part of this we should also get a better understanding of the following:

• How much granularity do we need in terms of libs/modules
• Where will the repo(s) live and how we will structure them
• How can we test these components in isolation and together
• How can we publish these components
• What does the component usage look like
• how to deploy code that uses these components
• What Node / NPM versions / environments can we support
• Can we support multiple server frameworks? for example Apollo Server supports express, hapi and Koa