Industry Session - Some pragmatic examples of Business model innovations on potential EOSC-related usage
22 November 2018 | 14:00 – 16:00
Austria Center, Vienna
Session drivers
The panel discussion was developed around 4 main topics, which were aiming to explore the relationship of Open Science with the Industry, the benefits of OS from an industrial perspective, the development of the EOSC and the strategy and goals of the EOSC.
The workshop that captures the spirit of how the “business model innovations could work through EOSC-related usage” was both timely & critical at this stage of the launch of the EOSC implementation roadmap, as it leaned towards the need of guidance and support from funding agencies and Member State intervention.
In the past twelve months, and since the previous SF edition, the “EOSC-intense community” has gathered relevant and substantial information from policy and project-oriented sources that it should draw upon to put together a stimulating programme with thought-provoking discussions of different stakeholders, such as Scientists, Developer Scientists, Service Developers, Service Providers, Infrastructure Service Providers, EOSC Managers. Different sources of information and subsequent results and guidelines should be drawn upon to ensure the “industry-related” session is both inclusive, involving all relevant EOSC players that also involve service providers engaged with the major research infrastructures.
As highlighted in the recent report by the European investment Bank published in July 2018 on Financing the future of supercomputing, there is a need of action on the part of EOSC for private industry suggesting that the business model of the EOSC needs to be structured around data, not primarily infrastructure. The unique selling point of the EOSC is the magnitude of data in the context of the convergence of HPC, Big Data & Machine Learning. Moreover, demonstrative use cases and data security are key success factors. The EOSC will be an incentive for private users if there are commercially viable fields of application that are highlighted in its usage. Similarly, security certification would also be a necessary topic to engage private users.
In this framework, the main objective of this dedicated to the industry workshop was to provide some very useful insights as to the ways in which the triple helix of industry, academia and the public sector could be activated and work more closely in order to produce interoperable infrastructures, to share data and IP in a secure and trusted fashion, to transfer knowledge and ensure that the best possible service is offered to their respective constituencies. Αn important step towards building a dependable open-data research environment where data from publicly funded research is always open with clear incentives and rewards for the sharing of data and resources could be achieved.
Image 1 - Preparations of the Session
Image 2 - The Panel
Main outcomes of the session
The session firstly focused on the importance of Open Science and its relationship with the Industry. As noted by the panellists, Open Science can create fruitful new relationships between Academia and Industry, assuming that the EOSC will foster the establishment of commons where data will be more openly available, and assuming there will be specific policy actions, like the ones already exist, to help people with smart ideas to leverage data which are open and use them to create something innovative. All researchers in the panel agreed that research is largely benefited by the immense amount of data that are available online through Open Science. However, Academia needs a strong relationship with the Industry to fully exploit its potentials. Researchers need access to the current and latest IT Technologies and innovations. If only academia is involved in this process, they would fund excellent IT infrastructure, but in 2 years’ time, it would be outdated. A hybrid model, with the industry involved will work better. The industry will also benefit as they understand how science is changing and in innovate towards that direction.
The value of Open Science for the industry can be multiplied. Open Science can help avoid the development of a different software or electronic infrastructure for the same purpose, by multiple actors within a market, especially if one of these actors is a publicly driven institution. Using open science can enhance the cooperation between these institutions and the companies in the field, thus creating a more effective infrastructure for the whole market and avoiding the waste of funds and productivity to the same goal. An added value is created also if Data, apart from available, become understandable. There are examples, like Copernicus DIAS platforms, that have already succeeded in creating new markets through this simplification of data.
In order for industry to take part and also share its scientific data and results, it will require them to realize that by widely adopting open software, their business will be increased. The way to do that is to make EOSC highly relevant for the private sector. There comes the question of how Open is Open. The federated access to data has huge benefits, but making it open with an expectation to get something in return is a matter of debate. There has to be a quid pro quo, in case of the data that is given. Open doesn’t mean that everyone is getting it for free. Thus, it would be interesting for the industry if EOSC provided also services that are being paid for. On the other hand, the private sector would also benefit if EOSC could provide insights about future needs and some standardization about all the services that are cloud related. Very few cloud services have a hierarchy, while scientists are somehow lost in the local stakeholder’s service providers. If this was organized and structured, it would also be very useful for the private sector, because standardization is a key factor for better and useful services.
The feasibility of building the EOSC within the next 5-10 years was also largely discussed. The need of involvement of the private sector in this procedure through private and public partnerships was highlighted. A key building block towards this direction is the evolution of payments. Vouchers could be a good starting point with the creation of a cloud credit/coin model as a next step for more flexible requirements. These coins should be used for the right data, data that are interesting for a global audience. Besides the technological part another important component is the community engagement with the creation of rewards and incentives system that isn’t at place in the moment.
The short term and long term strategy was the last topic to be addressed. On regards to the long-term storage of datasets, members of the panel expressed the view that a reference-architecture should be defined. This could be imposed in all the external providers that want to do business in Europe. Many companies are interested in long-term services, which could storage data for a minimum period of time. The opinions on how the services of EOSC should be charged differ. Industry representatives argued that most relevant stakeholders are looking for a subscription model, which will make a company sustainable. One-time fee may not be good for the project’s cash flow but there can be a hybrid more flexible approach for special services. It is also highlighted that the amount of the charge depends on the value of the data, the service and the nature of the user. The complexity of these models is not just to define a voucher, but how this voucher is exercised and priced. Finally the sharing and engagement of EOSC with other parts of the world should be made with proper safeguards, a strong mechanism of data sharing in place.
The Panel Discussion ended with some audience questions, regarding the risks of overregulation of the EOSC. Mr. Jurry De la Mar answered the question, highlighting that indeed, there is a risk of too much regulatory. Business is very agile and the challenge is to really be flexible and adapt, trying to really understand what the users need. EOSC would be wise if it didn’t try to think too far ahead, leave things open and discuss with the industry.
Main takeaway messages
Overall, the main conclusions of this interesting panel discussion are the following:
-
Open Science can create fruitful new relationships between Academia and Industry, if EOSC fosters the establishment of commons and data becomes more openly available, and if there will be more specific policy actions that will allow people with smart ideas to leverage data which are open and use them to create something innovative.
-
Data need to be understandable not only available.
-
Future needs and standardization of services included in EOSC will incentivize the industry.
-
Inclusion of paid services in EOSC could benefit the industry. Open Data doesn’t mean that everyone is getting it for free.
-
Community engagement with the creation of rewards and incentives system is essential for the EOSC Success.
-
EOSC Charging Strategy is a tough subject. Voucher System or Cloud credit/coin model for more flexible requirements should both be implemented in different kinds of data, services or users. The value of the data should also be taken into consideration.
-
There is a risk of overregulation of EOSC. Business is very agile, EOSC should not try to think too far ahead, leave things open and discuss with the industry.
AGENDA
Format
The workshop will begin with a welcome note and scope introduction by the Chair who will set the scene for the discussion. A panel discussion will follow. Each panelist will have 2’-3’ to introduce their organisation and make an initial positioning statement. A roundtable discussion will then be initiated, with the panelists answering a set of predefined questions. These consist of a set of questions to be addressed to all panelists individually and targeted questions to groups of panelists. The chair of the panel may also continue with follow up questions to the panelists, to further elaborate on a topic. The audience will be able to address questions to the panelists at the end. During the workshop a rapporteur will take notes and will wrap-up the most important statements at the end of the session. The workshop language will be English.
Speakers & Bios
D
S
Minutes
Intros
Giorgos: CEO of Lancom, the largest cloud service provider in Greece, stated that he only had the chance to learn about the EOSC project 3 weeks before the event thanks to the session organizers, while all the related service industry actors he knows, don’t know it either.
Rosetta: Research data management services in partnership with universities. OS is the basis for their roadmap. As an example all the RDM services are already integrating with OS Ecosystem and they aim to do so with EOSC too. Elsevier is also working in data sharing. They participate in the discussion as a service provider being keen on providing fair science data, to be able to integrate their services with EOSC.
Jan: EMBL has shared data very early on the 80s. Strongly committed in open data sharing and in making it available in fair form. OS is important for his own research in genetics. Co-leading a project where exchanging data with European investigators but also internationally. There is a lot of interest in OS. Support EOSC
Cristina: INFN, Dealing with the studies of Astrophysical. For this kind of studies they need cutting edge instruments and technologies. INFN develops both of them in their laboratories in collaboration with the industry. INDIGO Data cloud is one of the very successful projects in the software development core enabling data technology for research.
Jurry: T-System International, part of Deutsche Telekom. Part public-part private, have been working with science institutes and universities for years. As stated, they provide data and services to users and researchers. Users can do more and better science through Open Science, they are aiming in 2 goals through it: On the one hand, experience from projects to be reused and on the other hand, industry to understand what they can use from research that is relevant for them. A question was raised on whether finding better technology with the procurement rules that are available, or to develop new procurement rules to produce the most suitable technology.
Shaun: Chief Development officer in GEANT, it’s a pan European associate organization. Extensive differentiated network, with a range of services that operates across the world. They aim at facilitating communication between the scientists and the companies, and enabling scientists to communicate as fast as possible around the world so that science is speeded up. He argued that EOSC should go for quick wins and not only focus in the future.
1. Open Science & Industry:
1.1 How do you think Open Science will create fruitful new relationships between Academia and Industry, with a better flow of skills and ideas? (Please use practical examples)
Fabrizio Gagliardi: The current story in Europe is that there are a lot of good scientists and discoveries but when it comes to turn innovations to products, most of the times the best scientists go to the States to make their idea a business success. OS can create fruitful new relationships between Academia and Industry, assuming that the EOSC will somehow foster the establishment of commons where data will be more openly available, and assuming there will be specific actions to help people with smart ideas to leverage data which are open to use them to create innovation. Especially if the legislation gets harmonized. However, open data should not mean open to everyone. Good access not necessarily free, for researchers is free, but for SMEs maybe they shall pay a small fee. Eventually the infrastructure needs to become sustainable. If the SMEs grow, the fee should be augmented.
1.2 How do you think you will benefit from Open Science? As a research problem/solution contributor? As a user? As a builder? (Personal level & organisation)
Jan Korbel: His research is largely benefited from the fact that the type of data used within the research is data interesting in large scale. Very large datasets required. They rely on OS and communication of data sharing in order to put more data together and become stronger. It is also easier to replicate findings. Without OS this process would not work. However they are very careful and open to what they are doing as they are sharing very sensitive data across boundaries.
Federica Rosetta: OS is moving research in helping communities being more collaborative with the end result supporting more outcomes. This is the way Elsevier think in terms of roadmaps for its products and services. As a SP it is interesting to understand how science is changing and innovate, using their technologies. From the point of view of EOSC in particular, Elsevier is transforming from a publisher to a global information agency, focusing a lot in the data aspect and that requires Open Science, as they are doing it together with the research community. The success of EOSC is whether the community will get engaged. OS is an opportunity to continue the dialogue with the research community to be able to innovate in the right direction and create something that is really of added value for the researchers.
Doina Cristina Duma: The individual researcher has the ability to validate its result by a larger community. Another important aspect is resource management and sharing.
1.3 What is, in your opinion, the value of Open Science for Industry? Do you have an example of Open Science success story with Industry?
George Nolis: According to Mr. Nolis, everything that is open is beneficial for everyone and for the industry as well. But if it is not structured, it is like you search a niddle in the grass, so it’s very important to have OS and OD in general. Mr.Nolis presented a failure story regarding the absence of Open Science. In the beginning of their services they had a lot of software engineers who were working on a project. There is also a public company in Greece that was working on a project with the same final product/ service. So, following its completion by the public company, it gave it for free. If they knew that, there wouldn’t be a waste of time and money only to have, at the end, different software which does the same thing.
Jurry De la Mar: EOSC is making Open Science part of your day to day thinking, if you are producing science everybody knows it at the back of the mind. However things are successful if they are sexy. So Open Data also needs to be sexy. Their example as a Service Provider for ESA is indicative of this statement. As they stored and protected the ESA data, with the idea to make the Copernicus data free, the downloads increased. It made them develop a complete new platform, DIAS, it was so clear that this scientific data contains a lot of insight for agriculture, farmers or water management, etc. Like that, they say, they brought the users to the data and not the data to the users. With the example of Mundi, where they use Open Science they made data to look very simple, understandable, not only available. They hope that through EOSC they can find the next ideas that can be made to look sexy like this, and that the industry in other countries too, not only EU, would use and would derive insights from that. There is competition and there should be competition. EU is already so diverse, so are the applications, so there is not a single SP that can provide all.
2. Benefits of Open Science in General, an industrial perspective:
2.2 Do you think Industry will share its scientific data and results? What could incentivise industry to share? What is the best way to reward contributors to the EOSC?
Shaun Cairns: GEANT is involved with many international and local cloud providers. Their actual ability to engage with researchers is a lot quicker, their ability for research is institutional. Researchers quickly get access to their commercial services with a volume business discount, while on the Amazon/Microsoft side they get access to a channel that probably was very expensive for them to access through another way.
Fabrizio Gagliardi: In most cases, what counts is the business model. IT global vendors were making more money by using proprietary software until realizing there were money to be made if they adopted open software, because they realized that the market was increasing and they were selling services or other products. That is they key point, if industry realizes that by widely adopting open software and open more in general, makes them to increase their business, they are certainly going to do it.
2.3 How can the EOSC be made highly relevant for the private sector?
Shaun Cairns: This raises a question of how Open is Open. The federated access to data has huge benefits, but making it open with an expectation to get something in return is a matter of debate. Simple agreement with SMEs will be very attractive because they get access cheaply and quicker than they would. However there are IPR–related limitations for big industry actors, so there is a need to find an attractive option for these too. There has to be a quid pro quo, in case of the data that is given. Maintaining European ownership of data is a critical point. Open doesn’t mean that everyone is getting it for free.
George Nolis: It would be catchy if EOSC provided services that are being paid for. On the other hand, the private sector would benefit if EOSC could provide insights about future needs and some standardization about all the services that are cloud related. Very few cloud services have a hierarchy, while scientists are somehow lost in the local stakeholders service providers. If this was organized and structured, it would also be very useful for the private sector, because standardization is a key factor for better and useful services.
Doina Cristina Duma: The most interesting aspect is the technological transfer. Human capital too: industry may be interested in hiring researchers.
3. Opinion on the feasibility of building a European Open Science Cloud within the next 5-10 years:
3.1 The European Open Science Cloud is a very ambitious initiative. How long do you think it will take to have a significant impact on European Science, taken into account the fears that European Science is falling behind the US, etc?
Fabrizio Gagliardi: It is a question of scale, especially in Europe. Almost most of the communication operators they do offer some cloud services, often to rely on other infrastructures that is provided by the governments.
George Nolis: It’s mostly a political issue. It’s easy to defragment all EU sectors by issuing a guideline, that all the personal data should remain in each separate country. Like that, the local stakeholders will benefit. Of course we have to look on the challenges, especially interoperability because there are different aspects that have different demands. To make it feasible in 3 to 5 years we need to focus on public and private partnerships-no isolated efforts.
3.2 In the perspective of an incremental roadmap, block-by-block, which blocks would you implement first? Which are the necessary technical features to get broad adoption of Open Science by researchers?
Answered by: Shaun Cairns, Doina Cristina Duma, ,
Jurry De la Mar: It should be driven by value and by business. There is thinking on how we put an evolution to the payments. There is the model of creating vouchers, so that people can start using the service without having to sign a contract. For a large organization that has a requirement and would like to be flexible in the distribution of the benefits of a service, a sort of cloud credit/coin model can be created. That could be one of the key building blocks. It is very important to find the right data for these coins, insights or knowledge basis that are very interesting for large global audience. An example could be Genomics.
Federica Rosetta: Many issues: what would be the first step, how do you address the business model, what is the easiest for the user of the EOSC to approach the services, what is the easiest way to get to work. Beside the technological part another important component is the community engagement with the creation of rewards and incentives system that isn’t at place in the moment. Technically the definition of standards, and what is going to be the catalogue is essential.
3.3 Are you already working on the building of Open Science technical requirements? systems? solutions?
Answered by: Doina Cristina Duma, Federica Rosetta,
Jan Korbel: Researchers need access to the current and latest IT Technlogies and innovations. If only academia proceeds with the research, it will fund excellent IT infrastructure, but in 2 y time, it will be already outdated. A hybrid model can be a good alternative, with the industry involved. Very strong use cases will underline the added value and will facilitate to make the system out and running.
Shaun Cairns: GEANT is involved in 2 levels. They are about to start a new 4 year project which is really going to transform how the networks in terms of capacity performance and the unique characteristics that research organisations have for a network. In terms of the project, they are defining some of the business models for EOSC. The approach that they take is to integrate the minimum viable feature content that they can actually deliver. EOSC is to solve the pan-EU problem of procurement.
4. Short term vs Long term Strategy, Goals and Benefits:
4.1 What is the best way to fund the long-term storage of datasets?
Fabrizio Gagliardi: It depends on the business model. A reference-architecture should be defined, and this entire package could be imposed in all the external providers that want to do business in Europe. An option would be to demand that certain data must stay in Europe. If this happens and 20 years later maybe Google is gone, data will be able to be tracked; somehow they are part of a reference model that is understood. Certain data are needed forever.
4.2 What kind of datasets is the private sector willing to pay for? How does that affect the sustainability of the initiative?
Jurry De la Mar: The private sector has many options and many companies have long term service. T-Systems does it for Airbus, postal services, and the EU Satellite data for a minimum period of time.
4.3 How should EOSC services be charged for? Subscriptions? One-time fees?
George Nolis: All relevant stakeholders are looking for a subscription model, which will make a company sustainable. One-time fee may not be good for the cash flow. Of course there can be a hybrid more flexible approach for special services.
Federica Rosetta: There are 2 most viable options: Credits to access the services or getting all the founders to back data, by paying with grant money to access the EOSC.
Shaun Cairns: It depends on the value of the data and the service. For example genome data have commercial value. Different price will apply based on the nature of the user. The complexity of these models is not just to define a voucher, but how this voucher is exercised and priced.
4.4 How should the EOSC engage with science clouds in other parts of the world?
Doina Cristina Duma: Before going to other parts of the world we could expect from the EOSC to be a technical body to present all evolutions and solution in Europe. Moreover there are several existing groups, with other parts of the world. It is a good way to promote the standards and discuss what these standards and solutions are.
Jan Korbel: It should be shared with proper safeguards, a strong mechanism of data sharing.
Audience Question: Is there a risk that the rules of engagement will be too strict?
Jurry De la Mar: There is a risk of too much regulatory. Business is very agile and the challenge is to really be flexible and adapt, trying to really understand what the users need. Science is evolving. EOSC would be wise if it didn’t try to think too far ahead, leave things open and discuss with the industry.
Federica Rosetta: EOSC is an open marketplace, where different stakeholders can connect according to the terms and conditions. Different services will align. A marketplace is created and the dynamics of each market appear.
Jurry De la Mar: Open Standards are absolutely on the Industry’s interest, most of them are developed by industry. We live on them and it helps you to not only sell business next door but around the world, thus is very important. It is coming in the data business too.
LINKS:
INDUSTRY SESSION SLIDES: https://www.eoscpilot.eu/content/sustaining-european-science-cloud-some-...
All the sessions pages and the whole programme for EOSC STAKEHOLDER FORUM can be found at this link: https://eoscpilot.eu/stakeholders-forum/2018-edition/second-eoscpilot-st...