The big picture goal here is to choose an implementation per API, show how to launch it via continuous deployment on Azure/AWS/Google, and show that the APIs can be called together. And by doing this we give other projects a cookbook of things that 1) conform to known standards and 2) sample running code they can kick the tires on when learning how to interop with BDCat.
I think the problem is we can find partial or out-of-date implementations of most components (GA4GH APIs for example). Or more complete implementations (like Fence for AAI) but that has a lot of baggage and implementation decisions that paint us in a corner. We will have to pick the best implementation of each API and potentially fork and patch it to bring it up to date with the APIs definitions. And we'll need to figure out how to string together a minimal, API only (no GUI) set of services that can be used to find, access and process data. This is all in service of getting other groups to stand up infrastructure that BDCat will be able to interop with using these same standards. By providing them these guides, install process, etc we make it easier for them to understand how to interoperate with BD Cat in a concrete way (for data sharing, workflow sharing, compute sharing, etc)
The big picture goal here is to choose an implementation per API, show how to launch it via continuous deployment on Azure/AWS/Google, and show that the APIs can be called together. And by doing this we give other projects a cookbook of things that 1) conform to known standards and 2) sample running code they can kick the tires on when learning how to interop with BDCat.
I think the problem is we can find partial or out-of-date implementations of most components (GA4GH APIs for example). Or more complete implementations (like Fence for AAI) but that has a lot of baggage and implementation decisions that paint us in a corner. We will have to pick the best implementation of each API and potentially fork and patch it to bring it up to date with the APIs definitions. And we'll need to figure out how to string together a minimal, API only (no GUI) set of services that can be used to find, access and process data. This is all in service of getting other groups to stand up infrastructure that BDCat will be able to interop with using these same standards. By providing them these guides, install process, etc we make it easier for them to understand how to interoperate with BD Cat in a concrete way (for data sharing, workflow sharing, compute sharing, etc)