Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: middleware open issues notes, from memory

...

  • GÉANT OER update and ICT-22 call proposal - Peter Szegedi

    PS informed the TTC about a new call under the ICT programme (deadline - April next year), which is highly competitive, but relevant to possible continuation of the work that has been done so far in OER. At first, OER was a 9 month pilot that lead to a development of the aggregation engine and a portal, later picket up by GN4 - it is now a separate activity (SA8). It is expected to launch the OER portal by the end of the project. This area will not be funded in the 2nd phase of the GN4, but the service has more potential than a GN4 activity, so there is a need to look for funding elsewhere. PS is now working on creating a strong consortium, including commercials, libraries, universities that would create proposals - broader community needs to be involved. ICT is the closest opportunity to receive funding, however, there are other calls that are relevant and will be looked into. 

    VN suggested to get to know the reviewers and previously successful applicants. 

    VC: reminded to bring it to the attention of the GPPC.

  • TF-CPR update - Laura Dunford

    Task Force meeting was held in Cambridge in October. The members requested to have a clearer overview of other TFs and SIGs, on which LD gave a presentation. There was a brief discussion about a possibility that TF-CPR might have to become a SIG. Other discussions included Cloud Services (what is it that they need to know in order to be able to promote those services), raising the visibility of the training, maintaining websites and blogs, and the future of the PeaR community news (still generated in the old TERENA way).

    One of the issues raised was user outreach - Cambridge office described some of their work, which made it clear that there are too many projects for a few people to cover. TF members took a tour of the NOC to better understand what they want to promote - Connect magazine article about NOC inter-operation is planned. 

    Reporting mechanisms of the work should be more coordinated in the future. Some of the ideas: 5 min update, quarterly reports, thematic updates, etc.


  • REFEDS Update - Nicole Harris

    REFEDS SC had a discussion about the VAT issue - due to REFEDS being recognized as a separate entity from GEANT, the Dutch tax authorities require REFEDS to pay VAT retrospectively. That leads to the cut of 20% and therefore a smaller work plan with less MM. 

    REFEDS SC are trying to look for ways to bring more sponsors in - Jisc and CESNET agreed to contribute, but more financing still needed. There is a possibility to move REFEDS to Kantara - this is also where the industry is sitting. Idea has been circulated for a while, but that never happened. 

    PSz: Could GÉANT look into possible solutions? Perhaps change the status of REFEDS like GLIF has requested (for example, become a TF or SIG)? 

    NC: SC wanted more information why it was decided, but GÉANT only provides a service for REFEDS, it is a separate global initiative. No intention to be 


  • TF-MNM - Nicole Harris

    Decision has been made to let it run out next September, as the TF is effectively dead and the Chair has no time to do it.


  • Review of the slides on the Middleware issues - Peter Schober 

    PSc reviewed the work on open Middleware issues (action item 20150708-03). The issues can be split generally fall into 4 categories: "tired but did not work" (one of such attempts was community effort led by SURFnet, which proved to be not successful - perhaps the aims were not clear or it failed due to a general lack of interest) and "things that are alive", for example EduID.se and EduID.ch common idea to expand the concept of federations, including the schools and lifelong learning - that brings some aspects of provisioning and pre-provisioning. One of the proposals is to create a task or subtask in REFEDS (or similar).
    - tried but didn't work
    - given up on (for now, may revisit)
    - currently being worked on (authorisation, e.g. operationable def of "academic"; entity categories; harmonisation of identifiers, attributes, data release, etc.)
    - potential new/future work items
    An example of "we tried this but didn't go anywhere" was the application domestication effort led by SURFnet, which proved to be not successful - perhaps the aims were not clear or it failed due to a general lack of interest. As for "given up on for now" maybe federated provisioning could be named, even though (partial) protocol and software support exists. Newer projects like 
    eduID.se and Swiss edu-ID (expanding the reach of federations to also include schools/K12 and lifelong learning) have aspects of data aggregation and provisioning across federated institutions (though not across federations). PSc will possibly follow up with SWITCH and SUNET about extracting "lessons learned" from eduID-style efforts.

    For potential work items: Question of the right layer, e.g. institution (were we're not represented nor visible) vs. NREN. Multi-factor authentication area as The last category would be "things that we did not try" - those perhaps could be moved to the NREN level. Multi-factor authentication area is an example - can it be brought to Campus Best Practice? (Unless you're SURFconext where you can do things "in the hub"). Another idea is remote identity vetting for distant education could be a service, but it is hard to provide. Guest accounts is another problem , creating a framework that allows other academic institutions (located close to the distance leaners) verify identity. Guest accounts remains a problem for many (develop new technology or use existing one? Mention shutdown of the popular FEIDE OpenIDP per Jan 1st 2016) and proper authentication to mobile devices that is often used as the second factor (nobody is working on that. New issues are e.g. proper authentication to mobile devices (where entering complex passwords is hard and highly inconvenient). Their role as second authentication factor also has to be reviewed in cases where the mobile device is "all the factors" (i.e., it's being used as primary device to access resources, not as second factor to aid with secure authN on a PC/desktop/workstation).

    VC: Campus Best Practices are funded under GN4-1 and will not be continued in SGA2 as of today. A concrete proposal could be picked up by the GCP. New training coordinator has been hired to increase the outreach and training within GÉANT.

    VN: There is a need to discus these open issues somewhere. Vetting is for institutions and REFEDS is for higher up on the abstraction layer (federations), so no intention to discuss vetting there.

    TTC will pick up the slides from Peter Sc.


3. Topics of Interest 

  • GRNET work on OSS/BSS and ITSM integrationYannis Mitsos
    YM presented an overview of the architecture. End-users can order services directly. L2 VPN service example was given. Decouple the OSS from the BSS, user only faces the later (via the one-stop-shop) during provisioning. All technologies are open source, will be made available to the community as abstract/modular as possible. It is not a huge software. The one-stop-shop jump portal is a commercial service at the moment though. This concept has been taken from the telecom word.
     

    SIG-NGN could be a platform for discussion, we have to catch up with telcos. Lots of NRENs haven't looked at the service scaling, which is why they have not thought about it yet. Manual procedures do not scale anymore. One has to have an UI portal with all the information (snapshot of all services available to Uni CIOs). It needs automated procedures at the back-end though. MD-VPN could be an example where such automated integrated service provisioning should happen at the GÉANT scale. GRNET is running this service in Beta to show an example. 

    YM could possibly give a demo of the service to the community at the next SIG-NGN preparation.

  • Report on operational aspect of service provisioning across e-Infrastructures David Groep
    DG shared the Dutch experience on e-Infrastructures. Incident response and authentication are examples for common set of topics among e-Infrastructures. User facing organizations have to deal with AUP. Common AUPs are important. The mandate to work together has been there since 2014. The Dutch came together 15 years ago and tried different approaches in attempt to find a sustainable model. 
    For example, storage from SURFsara, ownCloud from SURFnet and AAI from SURFconnext plus the community of SURF came together and got SURFdrive. In terms of funding, the bill should not appear at the researcher because they start shopping then and create silos. Researchers raise questions regarding European Open Science Cloud vs. Global, closed for a while, industry research, and what is cloud? PRACE is not in that paper, and PRECE is not that coherent. Users will go shopping if they did not get something - for that it is important to engage with the users early on and be pragmatic. Bottom up approach continues to work.  Furthermore, writing the proposal is as important as getting the money because it builds your community.

  • ICT2015 report and the "new world order"Michael Enrico  
    ME informed the TTC that PPP and industry participation is being pushed by the new EC. Money is being shifted over to telcos, and especially into the well-funded 5G again. Integration and consolidation of e-Infrastructures, pulling out the research element and opening it up to new markets, SMEs. e-Infrastructures are going for EINFRA-22-2016 (part 2) Open Call programme - to do things together. Burgeno's video at ICT2015 is available. Spanish company has been selected by e-Infras (EGI, OpenAIRE, EUDAT, lately PRACE) to run the Open Calls on behalf of them. GA has to buy-in to this. About 50 beneficiaries could be funded and most of them can turn into real users afterwards.

...