 Hi everybody, thank you very much for attending this session on Business Models and Procurement. I'm Paul Rouse, I'm employed within the Jean organization based in the UK, my role as Chief Community Relations Officer with Jean, but here I was task leader within task 12.2 and in particular I had a legacy responsibility within the Jean organization for procurement, hence my involvement. As we go through the presentation I'll invite my colleagues who are involved in this project here to invite them to introduce themselves and as we've seen with the rest of the sessions in the EOS Hub Week, this will be recorded so please don't activate video or microphones unless we give you permission. In terms of the approach to questions during this session we're going to use the chat functionality within Zoom. We've structured the agenda to build in a number of breaks there so there's a few segments we'll go through and introduce an opportunity for questions. We'll keep an on time, may choose to take some of the questions rather than the end as opposed to that section so we'll just handle that as we go, but please if you have a question raise it through the chat functionality. Okay so what we'd like to cover with you today, a brief introduction to myself in terms of what this task is all about and then very much it's focused around the use of case studies and I'll explain why we're focusing on the use of case studies just shortly, but we're going to look at three particular case studies around the use of vouchers to distribute access to resources, the concept of virtual access as a funding mechanism within the Horizon 2020 program to enable consumption at three at the point of use and then a model where one entity can act on behalf of others to aggregate demand and provide provision resources services for use on the datasets. We'll highlight a few key conclusions and then at the end we'll just explain how this particular task is going to go forward because there's an opportunity for you to engage, provide some feedback and an input. So in terms of what the focus and activity by the parties who are involved in this activity and you can see the identity of those organizations involved is made up of three parts. The first part has completed now resulted in deliverable 12.1 and this was to provide an analysis of demand within the research education community environment for cloud services and resources to support research. It also took an element of time to focus on the business models how those resources could be provided. That's been produced then sequentially the intention was that being informed by these demand identification actual procurement activity would result and we would pilot different procurement techniques, outcomes to provide some purchasing frameworks that would be available for use by the community and then finally there'll be an appropriate opportunity to do some reflections, review and provide some feedback as what worked well and what could be done in further projects. Now as you'll all be aware from participating in EOSC activity or interested in EOSC activity there's lots happening, lots changing and there've been a couple of pertinent things that have happened that have impacted on our particular task. That's the formation of the EOSC governance and the working groups in particular the sustainability working group which has started to produce a number of documents, I think many of you would know them as the straw man, the tin man or I think it then indeed in due course the iron lady will appear. They provide input and insight into the legal vehicle how the provision of resources services for the community's use will be made available for the financing and the governance regulatory and policy considerations. That's very pertinent to the work that we're doing. Furthermore there's been a parallel project called OCRA, Open Clouds for Research Environments which very much has as its focus engagement with commercial cloud providers be they in the infrastructure as a service perhaps a platform or service but also Earth observation so there's very much active procurement work underway there. The net effect of this when we were deliberating our work and activity was in stage two rather than undertaking further procurement work which perhaps is already being met and satisfied by the activities under OCRA we would take the opportunity to review through the use of case studies what's happened out there in related projects to inform the EOSC governance and particularly the sustainability working group with this deliverable 12.2 which will be a written study looking at key experiences from work already done or work underway. Now in doing this I think it's a common expression that most have heard of this concept of the data lake so this is a repository all across Europe where the fair data will exist it will be identified how it be stored access is all under the scope and discussion of a number of related EOSC projects but particularly for us in our activity we wanted to think about how people access the data in this lake so we have constructed our document and it'll soon be available on Serger we'll tell you at the end how you can receive a copy of that or interact with us on that written outcome is we've added some perspectives we thought it's important to look at the use of the content of this data lake from the perspective of a researcher wherever they may exist an actual institution perhaps a university or in this concept of a demand aggregator and one of the case studies presented shortly will explain what we mean by a demand aggregator and thinking for what we've seen for early deliverables about the EOSC legal entity itself it may be a potential that the legal entity conducts itself as this demand aggregator in amassing requirements across our community for resources to work with the data sets and provision those for use by that community in one single place. So a little bit about a bit more of the content what we've taken and methodology in the approach so considering those perspectives that I've just described of the researcher and institutional demand aggregator the things that are pertinent to the procurement and business models is about how easy is it to access the resources and services so it's not about the data the focus of this it's all about the resources and services be they storage compute as examples we wanted to very much test and consider how the ambition of making those services free at the point of use could be realized many of the consumers of resources to work with the data lake will be subject to procurement regulations so the procurement directive is important as well as financial standards and financial standards here the one that particularly will come up you'll hear about so and I think we've heard about yesterday as well in another presentation during this week is around VAT sales tax EC funding Horizon 2020 rules is also considered to be an important test I like all good job descriptions and other activities we have a catch all other perspective as well so anything that we thought was particularly pertinent so with those perspectives and these tests these criteria we have identified three at the moment and we're working on a fourth scenarios or case studies where we have had experience across the community where we can look at these criteria and determine how effective these case studies have been in achieving these outcomes or against those criteria the aspiration being then that there are learning and key conclusions which I'll summarize at the end of this presentation can be taken forward by the Sustainability Working Group as they inform and further focus and develop the structure and operation of ESC so the four case studies that we're going to cover I want to look at vouchers virtual access mechanisms demand aggregator and one that we can't talk about today but very much is still in our in RI is to look at an arrangement where public public arrangements exist within an expectation of cost recovery so one public entity will sell services to another public entity and expect to recover their costs how that can comply with those tests that we've identified that's really the scene setter and end of the introduction just looking at the chat I can see that there are no questions at this point so that's working quite well in terms of timing and our agenda so thank you for your time and interest I hope you stay on I'd now like to hand over to two colleagues who are going to cover the case study around vouchers it's broken up into two parts because we've had experience of two instances of vouchers so for the first instance I'd like to invite Marion from CERN to introduce herself and share with you the findings there Marion please thank you Paul so hello everyone my name is Maya I'm working at CERN as a project analyst and as Paul mentioned I'll explain a bit more in the case study number one which is around voucher and I think I can exactly I can control the slides yeah and so we start I'm going to start to talk about the H&C Cloud project because this is where we began to explore the use of vouchers to provide access to cloud resources to research environments so just some key facts about the project to give some context H&C Cloud is a pre-commercial project pre-commercial procurement project that started in 2017 and ended last year 2019 with the procurement budget of 5.3 million euros it was led by CERN and it grouped 10 leading research organization in Europe you can see them on the map on the right of my slide and together those organizations put some money to to procure in this project so they formed the bias group of this project and H&C Cloud was successful it won an award for outstanding innovation in ICT procurement and from the Procure app plus so going more into details of the project in itself and the core activity of the project was to pilot a hybrid cloud platforms for the research community and those were provided in the pilot phase of the project by two providers that were competitively selected throughout the H&C Cloud project so those two providers were exhaust scale in a consortium led by RIA and T systems and what happened in the pilot phase of the project is that CERN as the lead buyer procured on behalf of the CERN organization that I mentioned earlier large scale services to test and use those two hybrid platform so that was really the core of the project now on a side or another view of the project a side activity was coming out of a decision of the project partners to dedicate some of the procurement budgets not to use the the platform at the large scale usage but to to procure vouchers so we're talking about roughly eight percent of the procurement budgets this give us about 400 000 euros that was procured as vouchers the project partners decided to do that for two main reasons the first one it was to lower the entry barriers at two cloud services to end users and especially for the long tail of science so to users that are not necessarily using cloud services before using the vouchers and the second objective was to provide services to the end users of those 10 research organization so that they could use services that are not necessarily available within the organization so I'm talking about different architectures such as GPUs FPGAs but also software such as TensorFlow things like this so looking at the process very quickly what we did is that together with the BIOS group those 10 organization we defined defined the characteristics of the voucher schemes so talking about the face value of the voucher validity things like this then ExoScale Antisystem provided this scheme based on those characteristic and they gave us some vouchers so that we could test the schemes the BIOS group made some tests gave feedback the feedback was integrated and then we started distributing vouchers to end users then we wanted to collect feedback on the experience so we distributed an online form to those users and based on this feedback we drew some lessons learned that we documented and the document is published on the nodo publicly available so I don't have time to go through all the lessons learned that we identified during this project but I'll highlight a few and as I said the all of them are documented and I did the link in this slide at the back but the first point that I wanted to highlight is that we really saw from the feedback that the user they need a simplified user interface when they access the cloud platform with their voucher and up-to-date documentation and trainings because not all of them are used to use cloud services they need a simple environment where they can navigate they were also very keen on having cost calculators because they wanted to understand what they could use which services they could use and how much with the face value that they had with their voucher then point that we identified is that sometimes even though the users were keen on using cloud services via the vouchers they couldn't because this was not compatible with their data management plan so this is something to to take into account and finally they were really keen on having a clear policies at the start about what's going to happen to their data once the voucher is exhausted and where they could store the data after using the voucher so they were looking for long-term data storage solution so those are four points that I thought was was to highlight from the feedback but what I really want to stress is that we observed that the end users and the users of the vouchers they found that the vouchers were really a good way to start using cloud resources so one outcome of this of the H&C Cloud project is that voucher scheme is a good mechanism to encourage new users to use vouchers it's maybe not the most efficient mechanism when we talk about sustained large-scale usage of cloud services but for a small project that I defined in time and for engaging new user it's very efficient so what we did is we brought this idea of vouchers in the Okra project and Paul talked about this project in the in the introduction so Okra started this year and it's a project to encourage to do procurement of commodity cloud-based services for the OSC so bringing vouchers as a way to introduce new users to cloud services was very relevant here and so just a few facts about the project before I go on describing what we did with vouchers in this project and the Okra project is laid by Jant CERN React Group interest ATIA members of the project and it has a procurement budget of 9.5 million euros and bringing this idea of vouchers based on the lessons learned from H&C Cloud Jant as the lead partner of the project ran a call-off in summer 2019 under the Jant AIS framework so under the existing framework in the Netherlands and the aim was to procure a total amount of 500 000 euros as vouchers the outcome was that three providers were selected to provide voucher there was one Azure reseller one AWS reseller and one independent service provider which is Cloud Sigma and once we had this voucher the aim was to distribute it to a new user to the long tail of science and to do so we identified three distribution channel which is Eurodoc the European Council for Doctorate candidates the MCAA which is the Marie Curie Alumni Association which is which groups and all the researchers that either benefited from or are currently benefiting from Marie Curie Grant under the European Commission Program and the EOSCAP so the two first programs first channels were coordinated by CERN so I'm going to talk about it in the next slide and then the last one was coordinated by AGI so I live I will leave my colleague Mati to talk about it at the end of this presentation so going on with what we did or how we distributed vouchers to the MCAA researchers and Eurodoc researchers we started by creating a survey in order to gather the requirements from those researchers we really wanted to understand whether the researchers were currently using cloud services and whether they were they wanted to use some if they if they were not using some what would prevent them to use cloud services so really to understand the full scope of the the full the overview this was also a one of the objective of the Okra project this was done very earlier in the project in quarter two 2019 so before the call of for vouchers was run the results were analyzed we had 72 valid answers and we documented the outcome in a document that is publicly available and I added the link there so you can have a look following this action of of collecting the requirements we started the distribution because we had the vouchers in hand as Jantran the call-off in summer and we started in November last year 2019 the candidates had to apply via an online form to to get vouchers and they were selected directly by Eurodoc and MCAA representatives based on selection criteria that were previously defined as of today we distributed 15 cloud sigma vouchers we are expecting to distribute more in the following weeks as we have currently called for application running and we got more application for cloud sigma vouchers so this number is expected to increase in the coming weeks we distributed five AWS vouchers and two Azure vouchers and those numbers are quite low for AWS and Azure vouchers as the distribution of those voucher had to stop in January 2020 so shortly after we started the distribution because there was a VAT issue in the procurement process which was affecting the distribution of the vouchers to end users in different countries and regarding the feedback we plan to collect feedback from the users we were planning to collect it already at this time of the project however this also has been delayed due to this VAT issue I mentioned so we will now collect it in quarter three 2020 and again we plan to document it and share it with the community. So now we leave the floor to Mati so he can explain more about this VAT issue I just mentioned and he will also give more details on the EOSCUB channel and what was done there. Thank you Marian so I'm Matti Heikkuren and in addition to working for the workplace 12 in the EOSCUB I'm also the Innovation Manager of the project so I think the sort of business models and so on they actually fit quite well together so in a way the vouchers and the VAT issue was essentially a bit of a surprise because the cross border of VAT treatment of the vouchers that was clarified I think understood first of January 2019 so it's quite the reason thing so in the end what came us as a bit of a surprise is that the place of consumption or in VAT terms that will be the place of supply is actually the country where the user resides not where the server resides and in a small scale pilot this is of course sort of additional unbudgeted effort for the cloud providers because they need to undertake the VAT accounting for each of the countries and track the information and so on and to make matters maybe a little bit more difficult to resolve is that the distributor cannot determine the place of consumption with the certainty I mean if you get an email from somebody from certain domain it is basically up to this person to tell that okay which country he considers to be a resident in so the I understood that the the actual service provider has certain certain sort of extensions in a way that you you can make reasonable assumptions but distributing distributed channels saying that okay this person is here doesn't necessarily qualify for that so this is something that in the end it's something that the cloud providers need to need to manage and budget for and and so on so then that is something that is sort of ongoing the discussion about and how to do this in an optimal way and of course if you scale up the number of vouchers then then the overhead goes down it was basically just sort of slowing down as Marian said in this sort of initial stage when you have a small number of vouchers and you're kind of trying to figure out how to do this with sort of additional thing that delayed the progress quite a lot but I think it's not necessarily an unsolvable problem but something that they're actually part of the work but it's still work we'll need to look into in more detail and then if you go to the next slide there was sort of couple of learning experience related to this yours cub as a consortium not the legal entity acting as a distribution channel and these were sort of similar thing nothing that is not unsolved that can't be solved but it's just sort of quite a lot of effort of analyzing the terms of conditions that the user accepts and to understand that okay are the yours cub partners also sort of covered by the same liability release as the other parties in this so we did some work there and it actually contacted the service provider and so on then when it comes to the user support we actually had two batches of vouchers one of them were sort of experienced users who know what to what to expect and with them they were sort of delivered as is so they would contact for the support for example directly with the vendor and then the other one which is through the marketplace so we actually did the or we have done most of the onboarding process and then you run into this interesting questions in a way that who is the service provider because it is sort of on one hand you have the yours cub help desk and use your cub user support being ready to provide provide sort of additional support and and and just sort of forcing this sort of slightly non-standard let's say service provider into into the system that again was something that's required a little bit of thinking but I think again something that once you have done it once I think it's going to be quite easy to replicate this and so on so so yeah I think this is sort of more or less the additional lessons learned that we are going to work on for the rest of the project time and come up with the model in the final delivery so that was more or less what we had for the vouchers if I understood so I see one question in the chat thank you Matthew thank you Marion okay so we've heard there that the vouchers have been applied they've had some positive effects and benefit but there are some challenges in how they're actually configured there's perhaps some overheads that need to be understood around VAT and configuration of liabilities in the contractual relationships that are involved by the links in the supply chain in order that a researcher the end user can consume resources so thank you we'll just take a couple of minutes to see if there's any questions I can see Mark made a suggestion I think we'll take that one offline thank you for your suggestion and there's a question here from Yuri Yuri says does this mean that EOS will not advise researchers to use AWS or Azure this doesn't reflect real use of cloud services by academia in research to vouchers have monetary value okay so I'll respond to that in the first instance Yuri absolutely that's not the case there's not a recommendation that AWS or Azure should not be used I can tell you and you're here shortly from Hilker there's millions of euros being consumed through those very platforms at the moment by research education users all across Europe so that's certainly not the recommendation from this report it was just particularly the providers and the engagement in this pilot in Okra for the first use of vouchers for stimulating demand in the long tail of science where we had a surprise to use the words of the presentation of the learning about a particular VAT directive an EC directive on the use of vouchers which traditionally the VAT needs to be accounted at the point of supply here in this instance there's a whole lot more administration around it being accounted by for the place at which it is consumed and those suppliers weren't in a position able to do that because they were engaged for a different basis so we had to suspend those two particular resellers there's to be absolutely clear there is no message in suggestion that AWS or Azure shouldn't be used and indeed they are available to many irony users already through the existing IS frameworks and through the Okra program we've just launched a new tender as well to refresh those frameworks make available these services and ongoing basis so that's that's hopefully answered that one very clearly typically the vouchers do have a monetary value Marion explained in the slides there HNCY cloud allocated one particular value for the vouchers that are available still under Okra we're sampling with different sizes of voucher value to get some experience about what the the best or the optimum value is for a voucher for a typical research scientific application to get the benefit for the researcher either they can fund a typical engagement as appropriate is 200 euros insufficient is a thousand excessive we're learning what that distribution unit value is through the project and if I may add Paul please I think yeah due to the fact that vouchers in the two project had a manager monetary value it was very helpful that the providers could provide cost calculators because the researchers really need to understand how much they can use with those 250 euros 500 euros as we don't provide the vouchers as an access to this type of VM for that time of that much time the cost calculators are really interesting here and really valuable Baron I'm going to take the liberty I think we can squeeze in just one more question at this section can you pick up the next question that's just come in as well please or Matty either one of you should I read it out yeah please can one voucher be used across resource providers or do you have to have different vouchers for resources from different providers so I was actually just typing typing replies so in the case of Auckland all the vouchers were always tied to a single resource provider yes and that was the case in it in Cycler as well thank you I propose we're moving to the next section of our presentation here we're going to hear from our colleague Sergio allow Sergio to introduce himself just a moment where we're going to hear about the use of the virtual access funding mechanism within EC funding agreements and how that can be used to best effect to deliver projects in EOSC Sergio please yes thank you Paul so I'm Sergio Andreazia and I work for the EGI Foundation and I am the head of strategy innovation communications and with regards to the EOSC project I am also the World Package 12 leader and so my presentation will cover three main topics so the first one will I will for those of you that were that are not familiar with the virtual access I will give a brief introduction and then I will explain about the experience using this financial mechanism is within the EOSC project and then I will derive a couple of learnings for the that could be considered for the European Open Science Cloud so what is virtual access virtual access is a financial mechanism that is part is defined in the context of the Orazontal D20 program specifically for the work program on the research and the infrastructures and it is specifically associated to certain calls so it cannot be used in any kind of course that the EC will publish the call need to clear need to mention it in an explicit way it is a complementary mechanism to another one which is called the transnational access and I'm just focusing on this because it is the one that was selected in the EOSC hub so the main goal of this mechanism is to stimulate the opening up of research resources or services that can be reached virtually via communication networks outside at their own user base so it can be data sets, it can be portals, it can be kind of services that in general are already being used by a research group or research community maybe they are limited to a country it was no group etc but as they there is maybe a wider interest for the European research the goal is funding to stimulate free of charge access to new users or to new users in general the mechanism as it is structured it was designed for services that in general that does not allow or do not need a user selection or do not need a user identification although user identification can be technically supported by the service for instance you can think about an open data set and who can use this mechanism so in the specific grants or calls where this is activated this can be used by the beneficiaries of the project so you need first of all as an provider you need to be part of the project what are these besides the obligations that institutions have as being part of the been a beneficiary in H2020 projects this mechanism adds some extra obligations if you want to use it to open up your services and specifically you need to publicly publicize the service in the context of the YOSC the YOSC marketplace is the main channel the service should be already TRL-8 and also there need to be justification and demonstration of the service also actively used by new users or new communities so there's an increased usage so there are metrics to be collected and reported today see through the project office and finally it need to be open the service to all European international user communities so what costs can be reimbursed so the virtual access mechanism that we adopted in YOSC would allow only to reimburse what is called actual costs and only the operational part of those costs so in general you cannot if you need to invest more on capital to increase the capacity of the service the capital expenditure could not be reimbursed so it's more only the operational part and only the costs that are really identifiable as those needed to open to increase the access for new users and I put the link in the slides of our presentation from the EC where you can have more information about this mechanism now I move to the experience of the YOSC so YOSC we are using the virtual access in four main service categories overall we have we are supporting 30 services provided by 42 partners with an overall budget of 3.8 millions of euro and these categories services are the common services like the more infrastructure layer like virtual machine management storage systems data management services and so on the more the horizontal layers then we have n-thematic services which are more the scientific specific services that were selected via an open core mechanism during the preparation of the project and here we can find I don't know high-level analytics platforms or science gateways or that usually are typically up per scientific discipline and then we have two collaborative services and that's an important category also we have federation services and here we need those services that are needed to enable the hub as a mediator between demand and supply in the context of the YOSC project so in the if we think about what's going on on a wider sense using the more known terminology nowadays when you hear about the federation core so those federation services are services that potentially are part of the federation core as I said the first implementation of the federation core so we have we find accounting monitoring services we have desks and so on they are not directly used to request by the users but they are needed to enable the european princess core so now moving more to the benefits what we saw as a benefit in using this mechanism first of all okay it was clear that even though only a small part of the cost would be covered the multi-operational cost of the marginal cost in core to open up services the mechanism stimulated the participation of new providers so they were early adopters early of the european official concept who wanted to meet the moves and also co-founded the service provision also it was good to boost up the marketplace with a wider service offering and it was the mechanism was piloted so we acquired the experience on how it can be used both for supporting services and for research and services and the community started to develop also expertise in using the mechanism for using the mechanism and also learning how to report so the measure kpi is like a number of users visit to the marketplace how many countries have been reached out for user satisfaction and also we collected useful experience for the european commission because one of actually the key points from the european commission was to enable this mechanism for the call that funded the osuka project was to learn how this could be used in the context of european princess cloud so with the initial report and the experience already they see took into consideration the initial experience and introduced the change for the next calls like the one the influx seven or three that for those of you that are familiar that have a deadline in june so they authorized the virtual access to use also the unit cost mechanism which is something already served for the transactional access and both mechanisms in future calls can be combined there's also some learnings on how this could be there are also added of improvements that we learned through the project first of all we found out that many providers did not join because the either the mechanism was incompatible with the national mandate for instance they couldn't subsidize the capital part of the investment they needed to do also it was difficult for them for those providing a paid service to adults a free access via virtual access because some adaptations were needed and there was no direct funding to do this adaptation so the scale of the service provision in the osuka was not enough for them to address the needy changes measuring also the measuring the growth of users for some services is not easy because okay the ec says okay you need only to claim the cost that you need to extend the access but you cannot claim your current cost sometimes it is difficult to have evidence of this the overhead for the project was much higher than anticipated so as a foundation according to the project we because we need to collect all the all the metrics from the providers and provide reporting to the to the ec so the low readiness from the providers was let's say a barrier to have a more efficient process and also the guidelines at the beginning were not so clear so this is also an area of improvement so my last slide is about the what what we can say for the european open source cloud in general we can say that okay virtual access it is a good mechanism to uh because first of all it focuses on providing three series of point to use and also to fund uh to stimulate opening up of services that is the goal of the osuka services and resources including data and the possibility of using both actual cost the unicost is welcome it will be very interesting to see how this will be used in the future course and it will be important to if they recover the full marginal cost will be possible including the capital expenditure that will stimulate more providers and also better guidelines are needed and also the providers need to increase the maturity and in also collecting all the metrics and maybe to automate the cost accounting and the distribution of the metrics to the policies and with this I would welcome any questions thank you Sergio we can see from from the chat window uh from participants we've had some interaction in this session so particularly to to Patricia I think she's reinforced with the comments that she's uh put on there Sergio that in her experience in the project called synthesis they've had some difficulty in understanding the mechanism around virtual access whether it's a local policy thing or within the actual rules so it reinforces from the EOSCUB experience of virtual access some of the experience Mark had raised the point that in the presentation the hyperlink to the power point doesn't work I think in this presentation now you're just in the visual image of the slide deck from the agenda pages for the EOSCUB week uh the slides were available there if you open them through that link then the hyperlink should work for you um so I believe that's just a technical system thing I don't Diego I don't I will double check the hyperlink after and I will fix it okay I think that's that's correct in terms of some of that Mark says he's looking at the power point okay well Sergio's committed to have a look at offline to make sure that hyperlinks working so you can access that thank you for the interaction thank you Sergio for for sharing with us our experience of how those relationships can exploit this new form of funding allocation within EC projects we'll turn now in our presentation to the last of our case studies uh I'll invite our colleague Hilker to present here around this concept we've termed the demand aggregator what could it possibly do for us in the EOSC world Hilker over to you please great thank you very much Paul so my name is Hilker Koers and I head up the data management services team at the SIRV in the Netherlands and I'm also part of this WP award package 12 around business models and procurement and I have the pleasure to take you through one of the case studies where we investigate the bits in detail the dynamics and the potential benefits of demand aggregation so I'd like to start with just a few remarks around the demand aggregation in general and then zoom into the specifics of the case study so the idea in principle is fairly straightforward if you aggregate your demands and you bundle your bound power you should be able to get better outcomes when you are procuring services the most obvious benefit I think too many is price discounts and of course that is a very important one but it's not only about negotiating a better price it's also about overall reduction of the costs involved in procurement there's a lot of overhead a lot of administration you save up on that if you do things together and also you can realize some service improvements not only in terms of the price points but also for example demanding compliance with a certain set of standards that are desired so price is important but it's not the only benefit for demand aggregation that sounds great so then you might if you would enter this territory new right you might ask yourself the question well if it's so and that all sounds great these benefits why is this not happening all the time and the answer is that well it only applies to certain particular use cases it needs to be scoped well and there's also some complications here so demand aggregation is most suitable for commodity type of service that are fairly standardized and that you can indeed easily roll up to an aggregate level if there's a lot of specialization involved this this method probably does not work well for you it also requires and I'll elaborate a little bit on the point through my my presentation that there are central parties who are really able to drive the process mobilize their constituencies and ideally take on a certain amount of risk so you need those parties in the middle to be able to do this successfully finally there are a couple of non-trivial legal fiscal regulatory aspects and also I'll touch a bit upon those in more detail in my presentation so why are we talking about demand aggregation here what is the relevance in context of the european open science cloud well it has been suggested by many that eos could potentially play a role of a demand aggregator or sometimes it's called broker we have to be careful with the terminology here so I'll also touch upon that in a little bit more detail in my presentation but for now just the message is demand aggregator or broker can mean different things so you really have to be specific about what you mean when you suggest that eos could be a broker or a demand aggregator notwithstanding the semantic confusion there is definitely potential here for eos to play such a role and then of course we want to learn from past experience about exactly what that role would entail and what are the opportunities and what are the faults so a brief illustration about well how you can think about demand aggregation and of course you can aggregate across many different dimensions I've plotted here the two I guess most obvious ones research domain or country in the case of research domain and the case study that I will be presenting is really about a generic service it's really commodity cloud services so bare VMs that can then be further configured to meet specific requirements but the service that's procured here collectively is a very generic service which means that you can aggregate across different research domains and of course also other countries and there it really helps to build scale if you're able to procure crossing the national boundaries and this has been enabled by this famous European council directive from February 2014 now the interesting thing about this particular case city is actually it does both because it focuses on commodity type cloud services infrastructure as a service and also it aggregates across countries so you get basically two-dimensional field of aggregation so a bit more detail about the case study from a high level um so we looked in depth at the uh the pan-european ea's infrastructure as a service standard that was run run by Jayant in 2016 and Marion already alluded to it in her presentation also connecting it to the voucher schema so it's interesting that these things can be connected it's not necessarily necessarily mutually excluded so started in 2016 the tender and the demand aggregation process was managed by Jayant together with the end ends the national research and education networks of which service one which is how service involved in this and the outcome is resulted in aggregated sales in excess of 26 and a half million with 385 institutions across Europe benefiting from that now to come back to Yuri's point you raised a concern about AWS and Azure I think they are represented here so just to underline the point that that's definitely on the table as well for EOS so the process again at a high level it began by preparing organizing and organizing here most notably means gathering the buyers to form a buy a group and settling on the requirements and specifications then there's the formal tender procedure which was managed by Jayant in their capacity as the central purchasing body and that resulted in multiple framework agreements then to actually start delivering and the service and the actual service commencement there is a step which is that the end-run and the service provider needs to sign a service commencement form and then the participating institutions which could also be the end-runs they can start to really make use of the offerings and that goes through a call-off either that's a direct award if the type of service that is needed exactly matches the specifications in the tender or a mini competition in case there are additional requirements functional and unfunctional and that still needs to be settled and then once the service provider is selected the institution and the end-run typically work together on the details on the provisioning but also to help support with training and adoption so zooming a little bit into the role of the end-run because we thought that was interesting as really an enabler of success of this type of scenario the end-run helps assures that the framework agreements are also in compliance with their national legislation this is of course facilitated by the EU directive but it's still not revealed so that needs to be validated and checked the end-run mobilizes his community such that they have an appetite to join this international buying group and also get engaged in in the discussions about the requirements and the desired specifications the end-run signs the service commencement form which is basically getting the green lights to start direct contact between service provider and the institutions and the end-run is a champion of the adoption of service through promotion education training and those type of activities now from a legal perspective the end-run can assume a variety of roles and so they have a choice in how exactly they want to sit in this dynamics and in this landscape of these different actors as a referrer the end-run is mostly behind the scenes organa or orchestrating mobilizing but does not have formal contractual role the other extreme is the undertaker where actually the end-run is procuring services from the supplier and then again delivering those to their constituency to the institute so they have a very active role which means they have more control more agency but also carry more risk and the reseller sits a little bit in between where there is a more immediate connection between supplier and user but the end-run still has a role to play also contractually oh underwriter yeah thanks Bob that's a silly mistake so underwriter not the undertaker i'll get that fixed apologies here it is correctly fortunately so in the particular case study that we zoomed in service net was the underwriter and in addition to their role in the process and the contractual etc they also created a portal surf cumulus which is a value-adding service that essentially helps to drive adoption and usage by making it easier for the the institutes in their constituency to get going and actually consume the service in agreement with the surf members it was discussed to add a margin to offset the costs that were made in both the current process but also in running this surf cumulus additional service so a couple of considerations coming from this case study that are important to bear in mind if we think about generalizing such a model to the eosk scale the tender process itself is subject to the legislation of the country in which the tender is organized the framework agreements the creation of those subject to legislation of the country in which the framework is drafted in the case of this giant tender there was the same but that's not necessarily so and then the call of contracts are subject to legislation of the country in which the research organization resides and also vat which was a thorny issue in the voucher scenario applies in this case in the home country of the research organization that's consuming the service so in this case that's very very clear cut so benefits realized was this actually a successful model unfortunately the exact detailed numbers on the reduced overhead costs are not available for publicly at this time however previous figures that have been reported to report cost savings of up to 25 percent or even 33 percent in Ireland these have been published before so I think at least that is gives you a little bit of a ballpark figure on the type of cost savings that could potentially be realized from from such a mechanism at scale so the key question is this a good model for eosk something that needs to be considered in more detail it needs to be more in mind I would say in spirit most definitely it fits very well with the spirit of the european open science cloud realizing benefits by collaboration at the european level there's potential for cost saving I think those reported figures underlying the point and again not only cost saving but also really driving the adoption of desired standards and it's if implemented well free at the point of use usually the central it department that pays the bill and the researchers themselves do not need to worry about that so that also fits with with what eosk is all about finally some lessons learned or considerations like I said in the beginning the process really works well for services that are indeed very similar commodity where the overlap in user needs and requirements is high if you're looking at very specialist services this is not going to work also lesson learned from the experiences with this geon framework and how that was deployed also including the work from my colleagues at service net is that the benefits realized and in particular of cost saves are the greatest when the research organizations that take part in this are really able and willing to commit a sizable proportion and you really need to make a fist or you need to have some some clouds in order to benefit from this type of cloud procurement the strong with the commitment to research institutes can give to the n-rent as an intermediary party the more powerful that n-rent can operate and make sure that those benefits are realized in this case we've also seen that strong national bodies in the process here have really helped to manage the progress organize it and mobilize them members at the national level that brings me to the end of my presentation I would welcome any questions or comments thank you very much she'll come I think the point that you're making about having a fist to make an impact for participants probably many will know that the size of the the IS cloud market was considered if you're sort of gardener data around three trillion dollars so for us to have an impact there doing something at a pan-european scale we've certainly got the potential and through this case study we can already see across Europe where we have aggregated multi-millions of demands that are going through those frameworks there's an opportunity for us to do something there one of the particular challenges that just like to reinforce and express that you picked up on there is a technicality around the mechanism for how this procurement is done where one party leads on behalf of a number of parties in the current procurement directive there's an obligation to identify all of the potential beneficiaries from that action that technically can be quite difficult a framework may last for four years if you aren't identified at the start of that that process then perhaps it means you're excluded and the activity of undertaking one of these procurement cycles is a considerable endeavour so it's not always easy just to add someone back into the list so to apply this at scale whether it be the EOSC legal entity in the future that perhaps does this on behalf of all users we need to to ensure that the mechanisms around the procurement directive allow all the potential beneficiaries who may wish to benefit from these EOSC resources can do so and that's supported by the legislation that exists in relation to procurement there is there is a sort of a follow-on question that Segway is into that from Carl Mello Mello asked about if Jean procured under the Dutch law he'll quote it if you want to take that or if you want me to cover that well you may correct me Paul if I answered this incorrectly but I think the answer is in this particular case now that was under UK law or more precisely I have that written down somewhere it's under UK law I think specifically the rules of England and Wales from the top of my head so for to build on that there's the UC procurement directive that's in place when the 2016 IS frameworks rewarded the procurement was carried out by Jayant in the UK at that time so the UK public procurement regulations were followed the contracts that resulted the frameworks were governed by the laws of England and Wales the current Okra procurement that's under way is actually being done under the laws of the Netherlands the procurement regulations of the Netherlands and the frameworks that will result will actually be subject to the laws of Ireland the call downs that are used by institutions, demand aggregators to procure resources on behalf of users they can be under the governing law of the the country in which they reside so there's a number of different parameters there there's the the procurement legislation the relevant procurement legislation that's enforced different member states that depends on who's carrying out the procurement or the aggregated procurement but the actual governing law of the contract can also vary as well okay thank you very much we'll carry on with the agenda now we're just going to move into the quick summary of the conclusions sorry apologies let me just pause a second a couple of questions I was talking so let me just check there okay there's one here very nice for the experiments of different procurement approaches from the three approaches presented which is more cost-effective in terms of investment against long short-term return which approach is good for which scenarios any comparison analysis results or recommendation okay team anyone like to pick that up the question is clearly clearly clearly scared some of our colleagues in terms of the structure of d12.2 at the moment and our way of working we haven't done a horizontal comparison of the the procurement approaches per se what we're looking in these case studies if you recall as I presented earlier is a number of parameters how it relates to procurement how it relates to cost eligibility how it relates to VAT and I think the reality is that for a successful and efficient and effective EOSC of the future whether you put that under what's the working title called EOSC exchange I would expect there's going to be a variety of these tools and methodologies applied depending on who the user is what the aggregation of demand is whether it's a niche service whether it's a competitive market how universal the service is so I don't think there's possible to necessarily draw a conclusion one which is better than another you have different tools for different jobs I don't think our report will will do a direct comparison that respect what it will do is identify the relative merits against those criteria that we've identified okay there's a couple more questions that are coming so let's pick up of these okay there's a question from Karl Maler here do you have any document of reference comparing pros and cons of these different national procurement legislation I can talk there from work that's been done at length and in particular from the experience of Gion when it legally restructured from a UK entity into into a Dutch entity and moving from UK public procurement law to a Dutch public procurement law whilst all member states incorporate or enact the EC directive on procurement each member state has a particular flavor and implementation of that so there certainly are some variances and we don't have an inventory of how that compares on a pan European basis for each member state we have some work in knowledge from myself and other involved in projects like okra of the differences between the the procurement legislation in Ireland for example the Netherlands and the UK but that's not being documented okay a couple more questions I think there's one that's come up here if it's been answered by subsequent chats any thoughts how to recover capital investments now excluded from the virtual access scheme Sergio I don't know if you want to pick up on that yes on this one okay partial improvement would come with the unit cost it has been introduced for the next calls and anyway it would be does need to be piloted with the next project like the inferior seven all the six subtopics are able to use that mechanism and we need to see what the experience will be because not all service providers are find useful or applicable to use the unit cost for instance some pairs like some feedback I got some more say I had some example of thematic services who said I need people maybe to operate the portal or to support users the new users so I need just pms and and that the unit cost is not applicable to them and so we need to see the experience and also we provided this feedback at the European Commission for further improving the mechanism looking forward to Horizon Europe so that would be interesting to see how this next round of experience will be reflected them in the next film or program thanks for doing I see Matty provided some some text response as well did anything you want to add further to that or happy with the text responses you're muted Matty anything you want you want to add further sorry it was for me yeah yeah I think this is basically for the for the follow-up projects I think this is something that as Sergio said we need to see how it works thank you okay we'll carry back on with the agenda is planned we're running just about on time conclusions then so thank you very much for listening to what we've shared with you in terms of methodology the approach for this task and some of the feelings or outcomes that we're included within this written report the work is still being produced at the moment we've come in towards producing a final draft and Sergio will explain how that's going to be disseminated shortly from this perspective of research or institutional demand aggregator we think there's some opportunities for embracing or ensuring there's that knowledge used by the sustainability working group other relevant governance activities within EOSC as the project progresses and we hope there's some value that can be incorporated there so these points there's lessons learned from the case studies and the application of relevance with regards to funding rules VAT compliance procurement legislation can be incorporated so they're designed into rather than learn from in implementation of course the things that we've shared here have hopefully stimulated some of your thoughts because we recognize this of course as we've seen from that the chat window the AII implications just as one example there's a lot of interdependencies working and we can't just look at this as a silo activity so the one slide perhaps to move on to here is just four points really in aggregate and summary I wanted to stress and highlight to encapsulate the key messages first one this is the task that we've been given is very much to look around the use of vouchers and how vouchers can be used as a value add mechanism to ensure the distribution access to research resources and services we've seen from the case studies this with some clear understanding tool in systems they can be a value where there'll be a value is perhaps not a universal application but there is a particular requirement with this VAT directive 1065 for anybody that wants to look it up that the VAT treatment works in a different way so the process and the systems required to record and capture that information and have systems that are capable of processing that volumes if we're dealing with long tail of science users so high volumes of users is a very much key characteristic that needs to be accommodated in respect of virtual access this option to ticking in your funding agreement to use it as a basis of cost eligibility to recover your cost to allow for an arrangement where one public body can sell services to another public body is in principle a good thing there's a fall on benefit for that as well that if the services become free at the point of use there is no procurement per se that's going on which can help solve any procurement directive obligations that the parties have as well now it's not to say that there are some challenges whether they're capex costs to how the services can be sized and dimensioned so there's a one case study that we've used there in particular around virtual access we've heard from chat window there are others who are using virtual access as a mechanism as well it certainly has a place as I described before we've visited this tool sets being used by eosk and it will hopefully have some application in particular the scenario of demand aggregation has positive benefits there's the IS framework that's been used now for a number of years and the expenditure that's going through those frameworks is one its way to 30 million worth of revenue to support institutions all across Europe so we foresee that as being a positive example of such an arrangement we just need to make sure that it's fit for purpose at perhaps a scale where maybe the eosk legal entity through eosk exchange or an actor within the eosk program undertakes this demand aggregation so we're have some petitioning working with the ec to ensure that we have the support when a procurement process is initiated and conducted that that description will hold scrutiny with the legal systems of all member states to ensure that those wishing to consume through those frameworks or those wishing to let the frameworks have a position of comfort credibility that the construct is compliant with those procurement directives the final one again here picking up largely on a procurement consideration we've got all the VAT overheads as well to to to consider we heard just day from the delinga project where VAT was at a point in transacting across borders how that system that process and overhead needs to be built into there but in particular where one public sector body wishes to sell to another public sector body if they're subject to procurement directives to procurement directives there are certainly regulations considerations or circumstances under which they can do that we have a space for such a case study we haven't yet completed that but that's that's something we intend to include in our report so before we move into next steps just to explain how you as a community can actively contribute access deliverable 12.2 before it then progresses into a final deliverable 12.3 Sergio will explain that that process to you we're very much welcome input just take a couple of moments now before we progress just to check if there are any other further questions that will come in if not I'll hand over to Sergio just to wrap up and describe the next step so let's have a quick look at the questions the chats it stimulated some discussion about vouchers virtual access some knowledge and understanding sharing for others there perhaps it's a first time for some reason the comments there that I've learned about this mechanism and what can be done so on behalf of myself and the presenters of the case studies thank you very much for your time and interest and interaction Sergio the floor is yours perhaps if you can describe to those listening what will happen next thank you yeah thank you Paul uh okay just one slide as a main summary to tell you how what are the next steps in terms of work package 12 and in terms of producing our results and making them available and also to give you an opportunity to have access to reports that you can comment on so first of all what you saw today is basically is the bulk of the uh content of what will be documented in the deliverable 12-2 and this deliverable is due today see in October anyway we thought about having an uh public draft version for the community to comment on and also for the usk system and the working group and we aim to publish this on the 8th of June so in the 8th of June there will be a public version and then we aim to follow the same mechanism the same process like done by the your secretariat there will be an online form with key questions and you will be invited you will have a time window to uh read the document provide feedback and then we will uh work on incorporating your suggestions for improvements for the final version that we give today see so if you want to this okay development of this document will be announced in the usk hub website anyway if you want to receive a direct uh notification feel free to drop me an email about this uh so okay this will be the most immediate step we have another step uh by looking forward to the end of June uh work package 12 and the task studio three of the project the studio three is about government sustainability are working on a joint collaboration because the sustainability working group had worked in the past on the fairing car and they've worked more on the why they're as a uh service recruitment and procurement models that could be applicable into the in the usk any various elements so we are gonna do a next step together and we want to do a joint work for which we will produce an internal report to the your sustainability working group and we are trying to align to the main your governance uh milestones uh so to make sure that what we do is useful to inform the governance on the main decision we take this year on how to shape the european open sense cloud uh so that would be the June uh step in october we will deliver the final output final version of the 12 to to the european commission and this will be available for everybody and then we have a third task of this uh work package which is mainly about evaluating your experience and defining recommendations for the way forward and so they will be separated deliverable for this we are discussing your scuba possible extension of the project of three months so the project is expected to end at the end of december we may postpone it until of march if we postpone it at the end of march this deliverable will be aligned with the end of the project we go to march otherwise it will be in december so the idea is to have more space to digest the experience also sourcing information from those organizations who are working on the preparation the project and which deadline is in june so we can incorporate all this learning for the next month into this final document and that's okay these are the main milestones and uh for this uh yeah we can summarize basically here we here you have the list of the team working on the 12 to the partners and uh as a lead contact if you want to have further questions following this work you can email me and i will pass the information to to the team so we'll hand it back to Paul in case there's any questions thanks Sergio no further questions at this time just check with the panel in case there's anybody that uh wanted to add anything further otherwise we'll uh okay in which case Diego thank you our our session here is concluded perfect perfect thank you everybody and uh yes now there will be the lunch break and uh you can come back here and connect uh after the break uh for the next session that will be uh uh 14 and 13 so i'll pass the two thank you thank you all for your attention we'll see over over lunch for a uh a chat about some of the things that we raised here the link naturally is the same for everyone so you can find it on the agenda on the USCAP