EGI-Engage ELIXIR Competence Centre teleconference
10/Dec/2015
Attendees:
Kimmo Mattila (Chair)
Gergely Sipos (note taker)
Miroslav Ruda
Fotis Psomopoulos
Kostas Koumantaros
David Meredith
Agenda topics:
1. Update on technical use case and infrastructure setup
- EBI OpenStack integration into the EGI Cloud federation
* First successful test of VM submission through OCCI
* Accounting and monitoring integration is the next focus
- ELIXIR IdP setup
* The IdP is 'officially up and running’? – TBC by Michal
* More than 100 registrations in the IdP so far
- GOCDB-ELIXIR IdP integration
* Technical work was done
* Compatibility with GEANT Code of Conduct requires further work: ELIXIR IdP will hash the ID before publishing this to GOCDB, so there will be no problem with storing and redistributing it (Hashing was proposed as a solution by the EGI-Engage AAI task).
* If AppDB and other SPs republish attributed (on GUI/API) then they also may need to comply with the GEANT CoC policy
- Operations Portal – ELIXIR IdP integration
* Gergely to collect operational processes for EBI, and decide whether these require ELIXIR IdP support in the EGI Operations Portal
- AppDB – ELIXIR IdP integration
* AppDB should support the ELIXIR IdP (it’s not the same as the EBI IdP which is already supported)
* Gergely to raise this to Marios
Dataset representation in AppDB
* Gergely to setup a dedicated meeting for this with Tony and Marios
2. Scientific use cases:
- 3 candidates so far. Info is collected about them at https://wiki.egi.eu/wiki/Use_case_candidates
- Marine metagenomics use case (Kimmo):
* Details about the required capaicity is in the wiki
* Need for deploying Hadoop Distributed File System or object storage, and Apache Spark in the EGI cloud. Gergely to check whether other communities already did this
* The question came up about resource allocation: How will ELIXIR or EGI (or together) decide how much capacity a user/use case can have? --> The CC implements pilot setup and resource allocation is something to decide later, for production setups.
- Phenomenal use case (Steven): Steven could not attend the CC meeting.
- cBioPortal replication use case (Miroslav, Michal):
* Further details should be collected about this, but what we know so far is that:
* It's a 'one site' use case, 'simply' hosting a portal on the site.
* CBio data set would be hosted on the portal server, user analysis jobs would run on the portal server.
* Probably only fraction of the CBio data would be hosted on the portal, because the total CBio dataset is hundreds of TBs, while the use case currently mentions 20-30 TB.
- We should collect information and details about the use cases in a standrad way. Gergely proposed to use the use case description forms that EGI is refining at the moment. These would capture not only technical details about the use cases but also details of the business case including potential user base and sustainability considerations. - Gergely to circulate this form to the CC list before the Christmas break, and the use case contacts from the CC get this filled with the use case providers.
3. Upcoming milestone (M6.3: Life science requirements analysis and driver use case(s) with implementation roadmap:
- Gergely to prepare a Table of Content of this in the EGI-Engage template and circulate this to the CC list before the Christmas break for feedback. Provisional outline:
* Scientific use cases: Descriptions based on the standardised form
* Technical implementation roadmap: Description of the ELIXIR Compute Platform and the required EGI-ELIXIR service integration activities (based on the document we wrote in the summer)
4. OSG role
No update.
5. Next meeeting:
- Kimmo to open a doodle about the weeks of the 11th and 18th of January
There are minutes attached to this event.
Show them.
