Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Brainstorm session of 1.5 hours on 21 August 2023 to generate ideas for setting the right agenda for the CTO workshop at end of September. 

Invitation

We invite you for a brain storming session on August 21st to prioritise topics to be tabled for discussion at the CTO Workshop (Above the Net) on September 27th. Infrastructure Cloud (including updates on the publication of the OCRE 2024 framework tender) will be the first topic for discussion at the workshop.

We realise that an ‘Above the Net’ strategy for the GÉANT community extends well beyond Infrastructure Cloud and thought to explore your thinking what other strategic services to include in the agenda for the CTO Workshop. These could include services in support of –

  1. Open Science:
    1. A collective approach to NREN participation in the EOSC (potentially as national ‘Nodes’)
    2. Institutional Digital Sovereignty
    3. Data sharing (Data Spaces) according to FAIR principles whilst retaining data autonomy (support for Data Access policies)

  1. Education enterprise support

  2. Other?

Recent conversations with the various NRENs at TNC23 and other fora have highlighted these topics, and we would like to unpack them regarding strategy and priority. Several of your NRENs already have related solutions in place in support of your constituencies and these could be very valuable in terms of reference to other NRENs.

Please join us if you can by accepting this invitation/delegating a colleague who may have valuable insights.

Notes

Present: Jakob Tendel, David Heyns, Richard Freitag (Sunet), Marina de Giorgi, Laurents, Licia Florio, Klaas Wierenga, Kurt Baumann (Switch, for Sebastian Sigloch), Aristos Anastisaiou (Cynet), Panos Louridas (GRNET) 

Notes: Jan Meijer


Main question: 


Klaas: bigger picture: we have this series of CTO workshops, that are supposed to cover whole area of what Geant community does in technology. Main purpose is to identify are there blind spots, are we on right track. Main purpose: are we working together on the right thing, the funding (e.g. the Geant project) is the secondary question. What are the things we're doing now that should be stopped, moved in different direction or started? What we found useful in previous years is a short (couple of pages) position paper that focusses the discussion. 

Licia: are there tools that are used in e.g. educational and how do we bring them in?

Kurt: am coming from NREN, so mostly regional, national but thematic nodes is in our focus. What we realised in last 2 years working on a Connectome project to offer research data to community by platform service we'd like to facilitate research process for community. All these communities are very heterogeneous. What we try to figure out is how we can find common areas of interest and how to coordinate these communities. Results of all these activities is that Switch is doing everything: everyone wants to help us, but does not have resources to do that. We have an EOSC mandated organisation in Switzerland (ETH). For us it's important we can clarify how is the impact of an NREN into EOSC, what is the understanding of Geant per-se, as a huge organisation they do all kind of stuff, they could help us understand things better, we could be more focussed on research data. Geant does a lot of things above net and now comes a new topic; research data. What I'm missing is some coordination to find the right way and do right thing. Solving local problems we do that, but have a better coordination would help us. 


Klaas: quick question: do you feel all NRENs are at same readiness level for that? Would like to find some common things most NRENs would want us to coordinate

Kurt: no, and we realise that. Several NREns do not have that mandate. If EC has designed some top-down approach for nodes that requires 

Licia: are you linking the above-net strategy to EOSC?

Kurt: exactly, but we realise (@Switch) that we have a kind of open research data strategy. 

Licia: should above-net include EOSC or not? Should role of (some, not others) NREN in EOSC be discussed elsewhere than in GÉANT "above-the-net" conversation?

Richard: methods of focusing in on "building the right product" and "building the product right" for research communities, possibly separate by domains but based on the same principles


Dave: are after services where the NRENs collectively want or must make an impact. 

Laurens: want to mention one thing which in NL is an issue: handling of privacy sensitive data. For that the role of AAI is of importance. Is an topic which is difficult for individual institutes. NRENs could play a role here. Another thing where NRENs could play a role is in supporting reproducable research. Have FAIR data, heading to reproducable software, but to reproduce research also need reproducable infra, e.g. storage and compute. Two topics which in NL but also within RDA community are of importance. 

Klaas: another topic I have been thinking a bit about and I know is important @ SURF: the public value. If we can come up with a set of public values and requireemtns we can put on e.g. cloud frameworks. Your AAI triggered that, we require any service offered through framework to follow federated identity paradigm. Could also be much more esoteric, e.g. human values type of thing. I've seen SURF successfully negotiate on likes of privacy. By having NRENs across Europe join forces we can be a sizable factor in these negotiations

Laurens: would be a good topic. In relation: a discussion about digital commons, digital research commons. Can be an instrument to safeguard these public values. In line of cooperatives, and the way Geant works. A digital research commons which has a goal to safeguard common public values, not only in our own organisation but also in 

Jan: can see how that could work, e.g. as coordinated NREN contribution towards the EOSC digital research commons

Klaas: I can see how this could result in practical activities. 

Kurt: and how it ties in with the nodes discussion in EOSC

Licia: see a defensive approach in nodes discussion "NRENs should be in", fear not all will be in, but you can't rely on past achievements, have to be active also now. 

Klaas: timeline for nodes is ... now, we need to focus to achieve an outcome on the node discussion that says 'open playing field' where you can join at any time, not a closed ecosystem where only few parties decide who gets to play. 

Dave: national node discussions, what NRENs are talking about is retaining a lot of autonomy for NRENs. Trying to identify what services NRENs would bring in as national node. 

Klaas: in mental model was thinking of AARC blueprint kind of model. 

Richard: is there a way to incorporate the non-participating NRENs directly from the beginning? We built e.g. SUNET drive as MVP we could hand over to institutions if so desired. Same can be applied to nodes, we 


Jan: summary: so common theme is supporting research, supporting research through a coordinated NREN effort towards EOSC, e.g. on national nodes, services. 

Licia: data spaces also fit in that basket. 

Klaas: and research commons

Kurt: agree

Jakob: previous to last summary is, when we're talking about NRENs roles as national and thematic roles was always intended to be in context of EOSC. Not every NREN has option of participating as national or thematic node. Should always keep in mind that having such an activity in Geant context, it needs to fit in a sufficient critical mass of the Geant members. 

Dave: focus seems to be shifting, focus of institutional CIOs seems to support to supporting research and science. Does make sense we're talking more about support of research from NREN and institutional perspective. 

Licia: unsure unis are shifting so much into supporting of research. Population that does research is smaller than student population. But more and more universities are outsourcing their IT to managed services. Means things shift to contracting, managing contracts, lock-in. Means it brings problems to other things outside of what cloud providers can offer. This we see a lot now. Unsure what we can do about this, whether we have the power to fix that.


Klaas: I think you underestimate the power of global R&E community. E.g. in T&I space multi-lateral identity federations is something Microsoft always said was not needed, but now they acknowledge this is  needed. But we need to join forces, we need to have a concerted approach, we need to understand what we want. We should then all agree "this is what we as a community stand for" and we require that from vendors. And if you don't live up, then no matter how good a fit you are for tech requireemnts, we won't do busienss with you.". 

Kurt: exactly, requires hard work. People, organisations, repositories, are willing to stay in contact with commercial organisations. E.g. we have someone in Switzerland working closely with ExLIbris. No chance to change things. 

Jan: summary: this is a variation on what above-net currently does but then for infrastructure-cloud: gather community and have a concerted dialogue with all sorts of commonly used commercial suppliers, not only for infrastructure cloud.

Laurens: second that, but starts with raising awareness. If you talk about values, should start with making that explicit. RAise awareness "these are our common values". When talking about research. Convince also decision makers at universities they have a choice. But if they choose to go with commercial companies who are not committed to safeguarding public values it's their choice. 

Licia: 

Dave: spoken to probably 8 NRENs who have developed own services with institutes, e.g. Copenhagen with DeIC data verse, a portfolio of data management services. Lot of work being done in different NRENs. If we can showcase what happens in different NRENs 

Dave: would it be good to have introductiosn at CTO workshop, e.g. Michel Wets on what happens in procurement, Switch on EOSC. 

Klaas: this was kind of a brainstorm and Jan summarised them. If we agree these are important themes to talk about I suggest we invite someone to give a short 10-15 minute introduction into these topic areas. Switch has lots of things in open science space they could present on, would be useful to have somebody to talk about SUNET drive, SURF about negotations with Zoom and Microsoft to have those themes there. 

Jan: agree, need to present potential scenarios for future above-net work, where the default is 'what we are doing now', e.g. focussed on infrastrucutre-cloud, and if you want a change in direction need to say so. And if it's an expansion of the work, we need to make sure we find resources for that. We can introduce the current direction, but would be nice with others introducing potential other scenarions.

Klaas: want our activiies to be value-driven, how do we add value for our members and your members. 

Panos: from universities lot of work in AI and machine learning. Challenges with provifding them with services they cant do themselves. REsearch communitites looking for accessible notebook-based workflows and backend infrastructure. Challenges with coordinating with commercial providers to supply this to universities sufficiently customised and on-time.

Jakob: might be good to have competence centre about dealing with how to use the offerings of commercial suppliers, WP4 Contract Management Group has a core competency here already.  Deepening our collaboration on using infrastructure-cloud. 

LIcia: would be good to have experience of Zoom/Kaltura procurements (Erik Kikkenborg)

Jan: have not yet heard about education?

Licia: found it hard to put fingers on this. Some NRENs are offering solutions for education, some others are more hands-off. All part related to supporting Erasmus but that was bottom up and unis have taken things in own hands. Something potentially radical: some unis have home-made solutin to exchange student records, others use commercial providers. Is an interest in these providers to let unis keep paying, and providers are not very responsive. NRENs could see if there was a place for them to either procure service nationally, offer a soluton nationally rather than going to commercial providers. Would mean stepping into market where many NRENs are not. 

  • exchange of student records
  • tools for doing education

Klaas: this overlaps with large scale pilots we do with wallets, using wallets to transfer this kind of information. Think that in education we benefit there is a large overlap between education and T&I space. 

Dave: academic credentialing and microcredentialling is similar challenge to health record sharing. 

Panos: about education: large scale pilots are about what individuals have in their wallet. The other is at universities: depts need to have details on programs, courses, at other universities, and their equivalent credit value etc. regardless of student coming. Falls outside scope of wallet.

Licia: would be interesting to know whether NRENs are really supporting universities or it's more a local matter. 

Kurt: following Licia's words, in CH we have AAI in place, students get ids, connect to services, access rights. Works. Each institution has their own local solution, are on their way to solve globally problem of solving problem of open research data. Our challenge is to bring all this focus together and think about how we can deliver a concept that can be useful for everyone. We're thinking about open source, not closed source. But many organisations have heavy contracts with commercial players we can't just replace. It's more or less recommendations they'd like to follow in this case, what Licia mentioned in beginning, how to solve this gap. But it's not only a procurement process to solve between different organsations, repositories, service providers, it's also a lack of concept. Domain space, data spaces are important in this view. 

Panos: more and more people are anxious and perhaps NRENs could have a role: AI and machine learning, see all these systems pop up and are trained on (assumed) public data. Not open source systems. We learn about the data that has been used afterwqards. Should we start having a say in how data is used in training AIs? Perhaps Geant could have a stance, as it is difficult to speak to Meta, Google, some coordinated action could have results in this. 


Jakob: needs to be outcome focussed:

  • what are current needs in community
  • are we well placed to deal with that
  • and is it relevant for enough of NREN membership to do collective action

Jan summary: different directions.

  • infrastructure-cloud (continue as is)
  • collective approach to commercial vendors of all sorts of services of relevance
    • evolution of current approach
    • potentially education tooling context
  • coordinated NREN contribution into research support (through EOSC), perhaps focussed on data
    • and include data spaces
  • AI as a new area
    • e.g. infrastructure and considerations for public sector trained datasets
    • e.g. interaction with commercial suppliers