Breakout: Federation, discoverability & interoperability#
What kind of things do we want to see e.g. share data between TREs, tighter integration between TREs…?
Why is interoperability important?
Standardisation across TREs in terms of how to do it is important. HPCs have similar problems to solve. Research questions to be addressed in a scalable way. Different issues, infrastructure, privacy, processes, can they all be addressed and standardised to keep from having to reinvent the wheel? This will make it easier to compare across TREs. Avoid continuing siloed approach.
What is the minimal amount of standardisation necessary to be able to facilitate discoverability?
We need to understand the data in TREs first before being able to address some of these questions (e.g. privacy).
Would like to see central management such as accreditation for TRE users (e.g. Scottish Safe Haven). Consistent userface to TREs. Governance is key to making some of this easier.
HDRUK alignment with data connectivity - expanding from COVID to wider subjects, avoid duplicating effort with what they are doing.
Interoperability (e.g., Docker as a standard) makes it easier for new institutions to engage with this and the extent to which collaboration can be achieved and deployed.
There may be challenges in getting researchers to engage in some of the tech required (like Git or Docker) so having consistency will help to reduce the learning costs for researchers to engage with TREs.
Need to answer big science questions and more pragmatically for data not to be locked behind artificial barriers. Makes sense to align and do things similarly across a region where data is similar - the transfer of code will make it easier to get the same study working in a new dataset.
There is a concern that we are just creating more silos - how do we manage to not do that?
Understand what is being used already across the research community? What tech? What data? Can we gain insight from the industry?
Make a list of the standards - understand the lay of this land and understand operability between them to see how these can be useful (e.g. operating systems and more granular systems like Docker).