20200226 - Meeting with ESCAPE, GW and DM TSP
Ian’s presentation
Resubmitted 12/02/01, will probably have some interaction with EU but none yet
Better technical overview
Name: Science Project, rather than Test Science Projects
“What is EOSC-Future”? A first implementation of the European Open Science Cloud (this is the only project funded under this call so it includes everyone → complex)
Interesting “who is who for EOSC” on agenda page (skipped).
Technical infrastructure: rather complex:

and in the technical part of the proposal it’s still under discussion as “composability”
Mark Allen: how do the ESCAPE services play into this?
Ian: rucio/FTS could be in the exchange layer already. Other things are a bit more complex (e.g. virtual observatory stays within astronomy but the technology may be of use by other people).
High-level checkpoints (at month 6, 18, 36) to try and tie the various aspects together.
Governance: the clusters (including ESCAPE) are involved in the “strategy and oversight” board as well as in the technical coordination board.
Task 6.3 in EOSC-Future is the TSPs, the work we did in the proposal appears here.

At the moment the emphasis is on the TSPs and the usability for others can come later.
Updates from the TSPs and discussion
Caterina/Elena: For now TSPs are “dormant” until we get the go-ahead on the funding and the project to start (but things are happening inside the individual collaborations). Timelines for hiring people and starting to work on this?
Ian: Discussion in the technical group. Potential of more feedback on the commission, lots of people who have to sign...May realistically before the project can start.
Next steps: make it clear that our needs are there in the technical specification.
Example: we would like to use HPCs to use ML.
Giovanni: we could think about : how do we exploit the ESCAPE service prototypes for EOSC? Do we need any further support or cooperation on technical aspects?
Example: credentials for users. But for us things may be different: what do you need to do your analysis somewhere? List requirements, have a “thought experiment” with the analysis platform/software catalogue/data lake.
We need to start immediately with the resources ready for the postdocs, and understand the boundary between EOSC Future and ESCAPE.
Steps forward:
-
Complete the table
-
What setup do you need for each TSPs? [uniform approach - Jupyter? Unique workflow? Make a short survey and interact with ESCAPE Kay/John]
-
What issues do you have in terms of data access. Do you want more than what is there? What authorisation is needed? Is it for a limited amount of people in the TSP? Giovanni can ask for more permission.
-
Can we plug the TSPs with the data lake? Can we produce pipelines for the ESCAPE data lake tests?
-
Catalogue (from WP5 in ESCAPE, Kay). The TSPs have to feed the catalogue with the algorithms/software and produce new ones.
-
The “credits” system: how to acknowledge people? There are internal activities (inside the collaboration) as well as transversal activities. We need to set this up.
-
Kay: technical implementation makes sure that contributors are nicely viewable and seeable by the stakeholders and by searching through the catalogue. How to make others aware, this is a question. But how to go both bottom-up and top-down?
-
Ian: Don’t forget about having papers on infrastructure on journals that accept them
-
Stephen: we can start communicating the TSPs with a political value and addresses national institutes
-
CD: need to inform ATLAS/CMS/EGO/big experiments beforehand?
Next meetings: will try on Friday mornings
There are minutes attached to this event.
Show them.