[Change]: Questionnaire summary: Reproducible deployments

[edwardchalstrey1]

Summary

The percentage splits of responses to the SATRE survey question on the possibility of exactly reproducing a TRE with the same infrastructure.

Source

SATRE specification survey responses

Detail

Survey responses:

20.1.a. Reproducible TRE deployments (e.g. through infrastructure as code)

No opinion	Not important	Nice to have	Important	Essential
10.48%	0.00%	20.00%	33.33%	36.19%

Where

TRE required features

Proposal

Although essential was only just the most popular option, we should add reproducible deployments as a required feature. We probably don't need to be too prescriptive in saying that IaC is mandatory, just that some minimal level of reproducibility criteria is met e.g. documentation that people can follow.

Who can help

Anyone

Specification section

https://satre-specification.readthedocs.io/en/latest/pillars/computing_technology.html (see #110) Currently missing

[crickpetebarnsley]

I think we need to be a bit more specific on what is meant by reproducible. The TRE will be a key part of the FAIR delivery process for research projects. It needs to be specified to help as much as possible.

A TRE should be able to be built and set up to the same spec and with the same roles / users / parties etc. Either if rebuild is needed from an archive or due to platform failure.

A TRE should hold all ran code so that any operation on the data could be sequenced and reran (assuming the data is (made) available). This is can double up as a audit log for actions and allow query of queries to identify policy breaches and risks.

A TRE should hold logs of the versions of software and the library configs used for each stage of work in the research project - it will help the "lab notebook" part of record keeping. Having a standard set of things as a basic level for a TRE will allow a baseline level of research record logging.

[edwardchalstrey1]

This is covered here: https://satre-specification.readthedocs.io/en/latest/pillars/computing_technology.html#deployment-management

Sone of your points @crickpetebarnsley are covered in that ^ and others in logging PR see - #132