Meeting on Thursday 2018-02-15
- David Antoš, CESNET
- Milan Daneček, CESNET
- Paul Millar, DESY
- Mikael Borg, NBIS
1. Updates and introductions
Mikael and I asked Nick (Car) whether he would be willing to become co-chair of this group. We are happy to announce that he agreed.
Now that the INDIGO project has ended, another different project has started that continues this work. This project is called eXtreme DataCloud (XDC). The RDA work in INDIGO is continued through XDC.
David described how he has tried to find some industry connection. This has not yet been successful.
Milan has joined this WG primarily as he will be attending the RDA P11 meeting instead of David. His main interest is in how this group can help with public tenders, in terms of understanding what different providers are actually offering.
Mikael's main interest comes from ELIXIR -- data storage & transfers. Unfortunately, at RDA P11, the ELIXIR session clashes with our session, so his attendence at our session will be limited.
Mikael described how he added some use-cases to Nick's pattern repository:
Mikael also reported that Sweden is planning to set up a national storage service. Our work here could be of interest to them. This service is at an early stage (no hardware has been procurement yet), but tentatively they will use CEPH and S3.
2. Updating our Case-Statement,
Paul explained how we have submitted a case-statement and was unsuccessful in gaining approval from TAB. There were some specific criticisms that we should address before resubmitting.
One criticism was that our group was very European-focused. We should demonstrate that the work has broader appeal.
To address this, we are very happy to have Nick as an active member of the group. We are still looking for membership from elsewhere (e.g., someone from America).
Another criticism was that the involvement of industry was unclear.
Although we haven't found an industry partner, under adoption, we can describe how different people intend to use this group's outcome in their procurement by describing the offerings and asking industry providers whether the description is reasonable. This may drive some within the industry to adopt this as a standard.
If this gains traction then it may be useful to adopt the work as an ITT standard.
Another related aspect comes from service discovery. Mikael described Consul:
This is a service that (amongst other features) allows service discovery, service metadata.
This is currently a back-burner project in ELIXIR, that will likely be most useful for their bioinfomatic users in discovering suitable comput resources; however, work is currently focused on extracting their requirmenents and needs.
The final criticism was on links to other RDA working groups.
Paul described how he has reached out to the various DMP WGs and the machine-actionable DMP description working group.
David suggested contacting the National Data Service WG. This WG provides a forum for people providing such services to discuss their work. Our work may be of interest to them.
As David has provided input to this WG in the past, he agreed to contact them and introduce our WG
Mikael suggested another possible connection: the Research Data Archetecture in Institutes IG. Mikael is a member and Ville is one of the group's chairs.
Mikael will contact Ville and coordinate an approach to introduce our group to them, to see if they are interested.
3. Preparation for our P11 session,
The proposed agenda seemed reasonal to those present.
Working through the agenda, we identified who is primary responsible and what work was needed.
The introduction round should require little preparation.
The update on our current status: as Mikael is (very likely) not there, Paul will do this.
The "Links with related activities" section is mostly an opportunity to talk with the people in the room from other WGs and explore the possibility to collaborate.
The "SKOS vs Ontology" section is a place-holder for discussing which technology options are available to us. Paul will contact Nick to check if he is able to lead that section.
For the "Service discovery" section: Mikael to look into Consol and Paul looking into Glue. We can extend this to also QoS discovery (given an endpoint, what QoS are available?). Paul can give an introduction on the work in INDIGO and XDC.
The "Plan for entering real-world data" section is offered to Milan. He could describe a plan for using the work to describe commercial offering and provide various "horror stories" as motivation for this work.
For the "EC ICT specification" Mikael knows someone who might be able to give us a brief introduction. He will contact them to see how it goes.
4. Next meeting
Paul wanted to schedule regular meetings for our group. He proposed meeting every two weeks: weekly would be too often and monthly too infrequent. This seemed reasonable to everyone there.
We should schedule meetings early in the morning, to allow Nick to join.
Paul will send a Doodle poll to allow us to choose a timeslot that people can generally make.