Ecosystem Approach Community of Practice: D4Science Business model proposal
From IMarine Wiki
D4Science is an IT initiative of a new scale and nature which requires innovative thinking and creativity regarding sustainability. From a recent international conference attended by Patricio Bernal, iMarine Advisor, only 5 or 6 international operators said that they were able to deliver and distribute this kind of global system today.
Sustainability is not a state to reach, it is transient. Sustainability concerns data flows, money, and many other aspects. Fund raising costs money; for 1$ invested, you can raise 3$ ! Need to go and talk to big players (from Serge Garcia, iMarine Advisor, first Advisory council meeting)
Sustainability can be defined from a number of perspectives:
Target of sustainability - from the preservation of datasets and their quality, to continued use and development of system infrastructures and their components, to the continuation and financial self-support of the interests and input resources of consortium of agents who constitute the D4Science/iMarine partnership.
Actions for developing sustainability – a comprehensive sustainability strategy needs to be developed, with indicative budgets, timelines and action points. Concepts and approaches for this include definition of a business model and related costing arrangements, identification of the market, and establishment of a legal entity. Based on this, and on agreed processes for development and sustainability, a more detailed investment and business plan needs to be developed and a governance mechanism needs to be defined and agreed, in order to ensure that the core areas of the programme can be managed and directed towards the fundamental longer term goal of the D4Science/iMarine partnership.
- iMarine has to set its own time frame, e.g. have a business model set-up by beginning of year 2
Options of business models applicable to D4Science
Range of basic options considered by Board 2
As preamble, it is worth reminding that the business model has to be adapted to the level of sophistication that involved players can afford (Source - statement by Patricio Bernal, first Advisory Council report). Various business models options can cohabitate. The identified options are:
1) A need for public funding support beyond the end of iMarine (2014) over at least two further cycles (ie at least 4-6 years): the consortium (or those who wish to take part) will have to negotiate ‘big picture’ science interactions at EU level and move towards becoming all or part of a global platform, with a financing basis increasingly derived from regional, national and development programme funding linked to the value of its downstream products.
2) The public partnership model: the partnership model is to be understood as a non commercial exploitation, largely based on co-funding through in-kind or monetary resources provision. Each community brings resources (data, software, hardware, administration, ...) according to its own expertise and primary interest. The Open Source software model falls within this category: D4Science led to gCube, which is open source. Compute (up to 5000 CPUs) , while storage resources are provided for free by external partners via agreements with private sector (MS, Engineering, …) that span over 2 or 3 years typically, just covering the project lifetime or a bit more.
3) Entering in major public-private partnerships for some or all of the core features of the D4Science/iMarine infrastructure. As examples:
- a current example is the MOU with MS for provision of cloud computing space), with sound contractual agreements. This could possibly be on a not-for-profit basis with revenues cycled back into the development of the system.
- the Global Oceans Observing System (GOOS): thanks for IT progress, it has boosted weather forecasts capacities, and was able to attract interest by oil platforms. Today there is a backbone of public services paid all over the world and additional inputs from private sector.
4) Developing private-public partnerships with little private companies interested by the commercialization (and possibly further development) of key components of the current system. Such relationships should ensure they did not compete with/diminish the function of the main system, and instead be based on negotiated arrangements for licensing or other revenue streams in order to continue to ensure the function of the main infrastructure.
- “Ocean related services” could be valued more. Close work with these potential clients would however be necessary to identify how they can make profit. Many potential clients and benefits, but there is a need to think out of the box. There are strong cases with the Business Cases chosen in iMarine, but somebody must have the vision to connect with “private users”. Developing such vision while proceeding with a market analysis could be matter of a project.
In this respect, EU’s policy to push private-public partnerships consitutes an opportunity. This idea is to catalyze private investments around a critical mass of public sector data providers such as ESA. It is expected that the potential size of innovative markets enabled by this massive availability of public data sources will attract private interests. The exploitation of these markets will require the development of multiple layers of software. In the future, EU funding could be geared towards helping Research developing prototypes, having them evaluated by the community and through the lens of agreed metrics after an 18 month period, and once approved, support industry’s efforts to make it operational through the professional services it is good at delivering. In such new logic, Governance mechanisms will be as important as technical aspects
The business model will have to take into account the different Data policies in various parts of the world, noting that it will be difficult to harmonise; here again, there is a framework of cooperating infrastructures among EU, USA, Australia which D4Science should exploit.
Proposals for Board 3 consideration
Public partnership model option
In the EU projects landscape, iMarine unique features include:
- the Ecosystem Approach scope, implying a range of Community Partners accross the Fisheries, Biodiversity, and Environment disciplins
- the Global dimension, enabled by the institutions of the consortium involved in the Board
- the strong IT component represented by institutions specialized in computing science and services
Each of these features can be declined in terms of values: (To come as a result of the Survey / iMarine SWOT analysis at Board members level)
The benefits perceived through the current collaborations with the institutions of the EA community can be threefold:
- of a "selfish" nature, ie viewed exclusively from one Organization view point; these benefits can arise from the ability to gain IT/KM services which this organization is not able to offer, or which the infra can offer in more cost effective ways. Among these services, some will concur to improve data quality, most likely through data sharing. Curation, or Taxon match, are two examples; any organization could benefit from advanced IT services which could be delivered on a fee basis.
- with a good understanding of the benefits achievable from data sharing and related derivative products while fostering through Partnerships complementary and synergetic roles; the Community of users and data providers (meant as Partners sitting in a Board) adds a Policy layer on top of what the infrastructure enforces as common behavior. Such community would articulate its commitments to one or few Product of common interest at the beginning (e.g. Code lists management), and each partner might make "updateness commitments" to a sub-set of products.
- The on-demand availability of resources (data, applications, computing) and a facilitated access to them.
The public partnership model builds on the above mentioned uniqueness and relies on the benefits which the concerned institutions expect to draw from a close and established collaboration. While seeking to realize the benefits, the partner institutions commit to contribute according to their capacities.
Community institutions can commit
- regular update of their data sources
- outreach, advocacy, and chanelling of project / funding opportunities to enrich the platform
- co-development of Open Source Software
- in-kind human resources to fulfill some role in the partnership activities (Secretariat activities, animation of working groups, etc...)
- fee contribution
- regular attendance to the Governing body
IT institutions can commit
- hardware resources
- in kind human resources for minimum software maintenance
- support to the identification, design and experimentation of innovative services;
- training support;
- secretariat resources for technical activities
In defining its business and governance model, it's worth noticing that the current consortium lacks some key capacities:
- a SME work force able to "sell" the iMarine platform upon funding opportunities, and to provide technical support/capacity building
- a SME provider of IT services having the commercial concern to serve the client as per needs expressed
- a legal entity (either existing, or new) representing the partnership and able to sign contracts and engage financial commitments
- a better and more strategic representation of "senior users"
- depending on the scope, declared commitments to GMES, EEA, GEOSS, EMODNET, MSFD, INSPIRE, FISHERIES
Various suggestions were made for sustainability models: A sustainability model across the value chain, in order to address the following aspects: Some aggregators become partners in EU projects, instead of the original data contributors. Other data providers, when interacting with large consumers, faced a costly license issue. Currently, the data is open, and it will not be simple to find paying customers for the data use, if that even is possible under the data-license. Donors pay for structure and secretariat, not for the scientists and the data. Data become ever more dynamic, and it is important for data providers to know where and how their data are used, also for them to prove their right of existence. If we can prove that, it will benefit all.
For infrastructure part (from representatives of http://www.europeanresearchfacilities.eu), these infras are partially nationally funded. They offer free services to research organizations belonging to the association and they get a fee from the other ones that use the infra (e.g. companies doing research experiments).
Ingredients of a follow-up project proposal
The D4Science/iMarine initiative is very innovative and its consolidation will require funding support from sponsors such as the EU before it can really sustain itself. During this process, each phase should see its Governance, business model, and sustainability making iterative progress. In other words, while addressing the consolidation of its business model, the initiative will seek support from the EU by applying to future calls for funding. What follows describes, with reference to similar developments in the EU context, the ingredients which should be feeding the next project proposal.
Similar to the ASSEMBLE experience, the current iMarine phase can be considered as developing the proof of concept and evidencing the demand. Having proved this, ASSEMBLE led to EMBRC which, on the ESFRI(*) road map, required to be partially self funded by national elements.
With reference to other experience reported by DG-Research and innovation, working towards a sustainable status, the set of Work packages aiming at defining the business plan should cover Legal issues, business plan and functional engineering.
WP Business Strategy: - Ensure sharing of common objectives and shared vision - identify main chapters of common interest, services, technical infrastructure planning
WP Legal work
WP Financial work - common methodology to cost and price the access to services Legal and Financial together should represent about one third of the effort, and aim at defining the Business plan
WP risk management and quality assurance
Legal instruments in the EU framework
(*) ESFRI: formal body which is research driven
ERIC: European Research Infrastructure Consortium, a legal instrument at EU level to facilitate the joint establishment and operation of RI of european interest. have a legal personality, qualifies as an international organization for purpose of VAT. ERIC are intergovernemental Organization operating under the subsidiary model. Ownership remains at local level. Core Secretariat with Staff seconded to the partnership Secretariat. Responsibilities of national elements (governmental commitments) versus EU part. Member can be Member States (min 3, and should have the majority of voting rights) and International organizations
Feedback from Board 3 and Advisory Council 2
Partnership model option
The contributions from community partners are primarily in-kind contributions, and the capacity to channel projects to contribute to iMarine developments.
iMarine current approach is very powerful and indeed you have to start from bilateral agreements (Carlos Morais). iMarine services can be used by the DG Environment Commission. You need to prototype and adopt a horizontal approach which can allow you to involve other communities.
Required functions for operating the platform
Maintenance and operation of the e-Infrastructure:
The D4Science e-Infrastructure consists of 680 software components, integrates a large number of heterogeneous technologies and exploits a considerable amount of data and computational resources. Currently this e-Infrastructure supports 16 VREs. Any sustainability plan should take into account these maintenance and operational costs.
Secretariat of the D4Science legal entity, will be in charge of:
- business development
- developing and maintaining MoUs with resources providers, or any private entity contributing to the operation of the e-infrastructure
- managing funds for co-development
- advertising/explaining the infrastructure capacity and services
- progressing policies for new business development (e.g. for a commercial exploitation of data available through the infra)
- Partners program management
- Services to users
- Data management, end-user relationships, provision of information about available services
- represent and ensure proper conduct in the e-Infrastructure and correct use of data products outside the eInfrastructure.
- manage the monitoring of user metrics in iMarine, for informing contributors on usage of their product, and for accounting and auditing purposes
CoP Board, and its supporting Secretariat:
In order to support collaboration, what is really necessary is the soft part, ie the human part shaping the community of practice. The role of the iMarine board is to facilitate the exchange of information with the communities so that the communities realize that iMarine is a network on which they can rely when they have a need for cross-disciplinary collaboration.
- ensure proper engagement of all stakeholders in the production, development and use of the tools;
- energizing the usage of the infrastructure; support the need to constitute groups of interest among the CoP that can approach the techniques and technologies that the e-Infrastructure can offer. We need to work in a very directed way and establish links with practitioners; to indicate what are the opportunities, to organize collaboration, and to train on how the e-Infrastructure can be used
- set-up and maintain best practices policy, in particular to ensure proper Data sets life cycle maintenance
Disposing of LOD engineering services:
Inputs at May 2013 iMarine workshop
Required actions towards sustainability
Ingrid Bergeret - 5 components needed regarding BC1:
- We need data, quality data consistency, etc.
- What services and tools are provided by iMarine.
- Governance model.
- A sustainable budget for a long-term view.
- Agreements to have data owners in the project, all member states involved.
Carlos Morais: supports this view that all member states should be involved in a coordinated manner, with reference to what is supported by FP7 and the GÉANT infrastructure example. In the fisheries context, DG MARE is to provide the policy component, including data management, workflow, data analysis, indicators for policy making and measuring, etc... which is key for engaging all member states in the same direction leveraging on this common policy and budget. An infrastructure such as iMarine can then support the MCS workflow, research, governmental and user community in a way that is very effective and very competitive globally
Mark Dickley-Colas: Whatever you build in terms of workflows has to be relevant to the people using them. Participatory approach, and flexibility across scales.
David Connor: beyond a technical and scientific exercise, you need to demonstrate that the applications and services work in a real-life scenario and what are the benefits, including at policy level. You also need to make the applications more appealing to and therefore usable by more people, especially people with less advanced IT skills. Showing the benefits of the products and services is very important. If you do not demonstrate the real-world benefits, it is difficult to fund it.
Amélie Knap: In the time remaining before the end of the project and the funds run out, the partners should try to get agreement with data providers and the different people iMarine is trying to reach to ensure their commitments to data provision.
Carlos: We encourage iMarine to take an active part in RDA. In Horizon 2020 there will be resources allocated for initiatives but they will be completely in the hands of the community. The idea there is to give the responsibilities to the data infrastructures developers. Communities aggregate around very focused topics and start developing the tools for the infrastructure in way that can be checked with the peers of those communities. RDA is global framework. The outcome of RDA is best practice guideline that the community agrees to adopt. This is a powerful mechanism for outreach
Rainer: One option is to create a consortium comprised of agencies which are both provider and users of data. Staff effort allocation could be connected to the benefits derived from using the services and software upgrades that can also benefit the wider community. Each would provide resources and allocate some staff effort to benefit from the services. Software is changing all the time, but upgrades benefit others as well. A group (ie consortium) can better serve others and also sell the services based on their value add. The other option is having a group with a shared interest. You already have the data, so a service can be provided, including the development of indicators. The EC would be willing to pay for that. A key advantage would be to submit a proposal to develop something in the iMarine area. Data provider agencies cannot pay but can bring data and use the system. It is important to identify now data providers and users, something between 10 and 20. It would set up a consortium based on a MoU, which is how FishBase operates, where FAO is also a member. Every year it decides on the policy.
Marc (concluding points): A model could be in the short term to shape through MOUs a consortium of core partners data providers that understand the benefits, and attract additional partners by offering on the demand side additional data quality improvement services (in the FLUX workflow for example).
Ingredients for a follow-up project
For Horizon 2020, the EC has decided on a list of five main priorities for their next work programme: 1. Data-centric science and engineering: Infrastructure for open access, management of extremely large research datasets, persistence and trust, as well as community-driven data infrastructures, and global coordination for research data. 2. Computational infrastructure: Support for the setting up of HPC Centres of Excellence, deployment of HPC Tier-0 services, support for open computing platforms and services. 3. GÉANT : Continued development and operation of the GÉANT infrastructure, support for international links and opening and strengthening innovation activities. 4. VRCs and virtual research environments: Supporting VREs as an open call (bottom-up). 5. Policy development and international cooperation: Global reach and connectivity; governance; sustainability; coordination with Member States.
Two priorities for data e-infrastructures have been identified: 1. How to implement, develop and operate e-infrastructures to meet the evolution of the explosion of data and promote seamless access, use, re-use, and trust of data. 2. How to facilitate the transition towards Open Access for Scientific Information and this is linked to researcher electronic identities as defined in the ERA Communication.
In the coming 6 months, you should be very active in preparing what is your medium and long-term approach so that we can get the best out of the first call in H2020. Before the end of the project you should undertake a cost-benefit study.
Data policies in support to sustainability
How to deal with the segmentation of the duties between different participants in the ecosystems. The segmentation is very important. You said that the initial validation ultimately lies with the data provider. In my experience, collecting data from different sources is the most difficult part to finance, but it’s a significant amount of work. The weakness of the system is actually quality control and validation. This segmentation will also allow for eventual policy changes that could be useful for sustainability
It's important to give feedback for the data contributed by original providers. But providing a feedback mechanism that works is really difficult. We need to make the reporting as easy as possible because people do not have time anymore. The process to report an error has to take less than 1 minute otherwise people will not report. The tool for pinpointing mistakes should be one click The real measure could be the number of people finding a mistake and see how many people report back. This could also be an incentive for a data provider to participate in an initiative. Why should a data provider share its data? The return of investment could be checking the quality of the data. This goes back to provenance and metadata, and how to automatically trace the process. This would be the first step.
We do not want to charge open access, so the user has not to pay for them. If 10% of the proposal goes to the data provider, this is perfectly enough because we are in more than one project
RoI of acting as data provider implies benefits through value added services; Data infrastructures offer new opportunities to create new data, so this type of opportunity needs to be acknowledged in the definition of the Business model with proper identification of related functions.
A private organization, marinexplore.org, has started to provide aggregated value-add products for marine data. They are stimulating data providers offering rewards.
Inputs to Market analysis for Board 2 consideration
Depending also on the outcomes of the business model, market characteristics of areas of particular importance for capital funding and income streams may need to be determined in further detail, to understand more clearly what would be the options, and how best the D4Sience initiative should be positioned.
At a fundamental level, the “Sustainability and Exploitation plans” deliverable of D4Science(-I) has described the potential market for D4Science as offering data, data-tools, services and a software platform.
The market analysis could include a number of actions, including but not limited to:
- Further assessment of trends and potential directions in high level science policy frameworks (GEANT, ESFRI/Emodnet, DG-Mare’s FCP, EGI, FAO-EU, Data rich Policy processes such as CBD/GOBI, IMO, International Seabed Authority): reviewing opportunities, competition or possible conflict, defining the best ways for D4Science to interact/develop/gain resources
- Specific opportunities such as participation in ESA’s proposed public-private strategic partnership for e-infrastructures and the EU-FAO strategic collaboration framework
- Options for commercialization with major private infrastructure/service groups (Eng?, or big player such as Microsoft, Google) ; key advantages/disadvantages
- Specific ways in which VREs or other products could meet emerging needs/create new market potential for clients (eg providing new functions/services which could add significant value to clients’ own offerings to their respective markets). This process could be carried out at various stages of the development of the D4Science system, but some of the outline characteristics/issues could be defined within 12 months.
- Promoting the “value of data”: data collected could be considered as marketable good. Market specialists should undertake a real economic approach based on the real value of data. Following are some examples of business opportunities regarding value of marine data: a business model could be based on encouraging data provision rewarded by access to Cloud time computing e.g. offered by Microsoft, or by access to free publications e.g. offered by Elsevier, etc. in exchange for rights to using this data. Some projects are funded by private sector that could use the iMarine data and outcome, e.g. : a Russian company finances a project 2M$ per year to avoid route of mammals (e.g. whales) when exploring for oil and gaz; 10000 drifting FADs in tuna fishing can provide coordinates, biomass estimate through sonar, temperature measurements.
- Exploring the market of improving quality of data collected: addressing the accuracy of data collected, and the role of iMarine data infrastructure in revealing errors and help reducing discrepancies among sources. Such role is a real asset for the infrastructure which should be dealt with.
- Exploring the market of Facilitating access to historical data:
Inputs to Market analysis for Board 3 consideration
The Market/Business opportunities listed below stem from the recurrence of messages on the demand gathered throughout all the documented collaborations This list is provided for triggering discussion and feedback.
- Market / Business opportunity 1 - Support to implementation of FLUX related data workflow: a suite of FLUX workflow services, including production and scientific exploitation of aggregated statistics and indicators
- Market / Business opportunity 2 - Support to access to taxonomic, biodiversity, and environmental databases and related analytical and environmental enrichment services
- Market / Business opportunity 3 – Generating a EA Linked Open Data network in support to the EA
- Market / Business opportunity 4 – Collaborative reporting tool, Content management support
- Market / Business opportunity 5 - Provision of "unilateral" data quality improvement services
- Market / Business opportunity 6 - Collaborative research
- Market / Business opportunity 7 - Strengthening sustainability of existing infrastructures by pooling resources
- Market / Business opportunity 8 - Provide facilities to other relevant infrastructures, e.g. ESFRI ones
- Market / Business opportunity 9 - On-demand provision of facilities to scientists of the public and private sectors.
Feedback from Board 3
- Market / Business opportunity 9
In the scientific community, many are struggling to run models, and they would be very happy to get easy interfaces. It would be good to be able to advertize how easy/difficult and at which cost a new model/algorithm can be enabled.
Ecoknows is another example of the modeling support which iMarine can provide: bayesian statistics to complete length weight relationships
Feedback from Advisory Council 2
- The set of collaborations articulated around Business Case 1 provides a very interesting landscape of collaborations which should attract donor’s support.
- some iMarine business lines are naturally supported by FAO, others by OBIS/IOC. Some unsupported lines may need outsourcing to private companies such as Terradue. The business plan should be articulated along these lines
- size each opportunity in terms of effort, time, money needed and size of interested audience, then prioritise based on these numbers
- on opportunity 1] iMarine could centralise/federate the set of developers required to transform xml sheets into database in each European country. Also, harmonise the processes for data presentation, so that sets can be compared across countries
- assess the willingness to pay. Would FAO be ready to pay ? If FAO says “it’s backed by UN funding for 10 years”, then this will gain trust. This belongs to talks about partnership model. We would have to ask our Member States.
- with regards to business model and funding, we play with Open Data, we could get money by arguing that we support this strategy. Push on support to government-funded initiatives that promote open data. Identify companies that sell marine data to governments or industries, offer our support
- Considering GOOS model, we have to be aware that GOOS worked mostly with weather forecast; climate is operational (day to day) and strategic
- For Fishbase, this platform would be a big attraction
Inputs to Market opportunities at May 2013 iMarine workshop
Assisting Inspire data providers to comply with Inspire directive: An e-infrastructure like iMarine could help to improve that, through provision to the concerned Node of monitoring and service support. iMarine could also be provided as their informatics capacity: hosting of geospatial data, transformation services, Metadata generation, ...
Among services provided to scientists, the capacity to reproduce the experiment/data processing is certainly a nice service.
Another very practical aspect is the classroom usage. To run an experiment it takes 3 minutes per person. The experiments are on our server and they are executed sequentially. It means that it takes 40 minutes to finish it. So there is a very direct need for a service where we can run classroom experiment in parallel exploiting your servers. It’s really a classroom usage of an existing model at the same time. Moreover we have to know how to connect easilly to the service. The demand for classroom services is rather for communities at the border of scientific community. The scientists are aware of the complexity. For the scientific communities you really need to sit down with them and explain what you are doing and they need to learn otherwise they will not come. You will have a larger market in those users at the border of scientific communities that try to apply, like NGOs etc
in DG MARE's context, iMarine could ensure that all controlled data goes to the scientific community. Secondly, in the controlled world, data is required that is in the scientific world. As the controlled world is unaware of the data types that can be provided (e.g. water depth estimation) it would be useful to provide selected services to the controlled world which can facilitate their work.
Think of ways of being attractive to potential customers:
- Quality of data is one of the streams on which you can develop new tools.
- Provide new data services along with the development of indicators could be funded by the EC
Communicating the aims for the programme convincingly is a priority, the initial stages of which need to be considered even ahead of a detailed costing and investment plan: a capacity to convey the motivation and vision to senior managers who will eventually support the necessary investments is fundamental. In this respect, at an early stage, a convincing vision statement for the platform and targeted messages to senior managers would need to be developed, and this would be followed through as the concept develops, with communication becoming a key attribute of the system and its evolution. A complex initiative such as D4Science concerns a diversity of stakeholders from software developers, through to data managers and statisticians, scientists and science communicators, and finally policy makers. These are in their respective area concerned by a range of D4Science Products and Services, and senior management decision-making will be influenced by buy-in of the various levels of involvement. Here again a communication plan should develop the right messages to the right clients. However care will be needed to ensure that these are based on customer needs, at whatever level, rather than simply trying to sell D4Science concepts simply because they exist.
A first draft vision statement was presented to First Advisory Council in March 2012:
<<The vision of iMarine is to provide an innovative and sustainable web infrastructure that will support the challenging, cross-disciplinary needs of the Ecosystem Approach to Fisheries Management and Conservation of Marine Living Resources, and facilitate the emergence of a more unified Ecosystem Approach Community of Practice (EA-CoP) through interconnecting these actors around common data, information and knowledge-building>>
Advisors were of the opinion that this was more a set of objectives than a Vision. A vision would need to be established at higher, more global level.
A second draft vision statement was presented to the General Assembly in October 2012:
1] <<By enabling integration and management of data and software resources across information system administrative boundaries, iMarine enables the actors at all Community of Practice levels (from software developers, through data managers, to scientists) and across various disciplines to collaborate in cost effective ways, for the delivery of enhanced science-based information products which serve ever demanding policy goals of the Ecosystem Approach>>
For Advisory Council 2 (March 2013), two additional alternatives are proposed which put stronger emphasis on the CoP aspects as envisaged in the original vision statement:
2] this alternative positions the role of the infra first, introduces how it supports policy goals, finally describes the impact on the building of the CoP:
<< The iMarine e-infrastructure will support a cost-effective and facilitated retrieval, access, collaborative production and sharing of science-based multidisciplinary information serving the ever demanding policy goals of the Ecosystem Approach. Through interconnecting all concerned actors operating in different domains around common data, information and knowledge-building it will facilitate the emergence of a more unified and effective Ecosystem Approach Community of Practice (EA-CoP). >>
3] this alternative places upfront the final policy goals, and stress the role of the infra in supporting these, and the impact on the building of the CoP:
<< Richer, better quality and timely science-based knowledge will be available to serve the policy goals of the Ecosystem Approach through the iMarine e-infrastructure. It will enable a cost-effective and facilitated retrieval, access, collaborative production and sharing of information and tools. By interconnecting all concerned actors operating in different domains around common data, information and multidisciplinary knowledge-building, it will facilitate the emergence of a more unified and effective Ecosystem Approach Community of Practice (EA-CoP). >>
Feedback from Advisory Council 2
The above alternative of vision statement were discussed at the Advisory Council 2. The third version above was preferred, slightly amended, and approved as the iMarine vision statement as follows:
<< Richer, better quality and timely science-based knowledge related to aquatic resources will be available through the iMarine e-infrastructure to serve policy development and implementation towards goals of the Ecosystem Approach. The e-infrastructure will enable a cost-effective and facilitated retrieval, access, collaborative production and sharing of information and tools. By interconnecting all concerned actors operating in different domains around common data, information and multidisciplinary knowledge-building, it will facilitate the emergence of a more unified and effective Ecosystem Approach Community of Practice (EA-CoP). >>
While discussing this statement, the properties which should prevail for a mission statement were proposed as follows: - the mission should fit into the objectives of many sponsors; we need a generic mission statement that can be adapted to different audiences - the vision statement needs to be kept very simple and generic - the policies keep changing, but the vision needs to be long-term and stable - this mission statement is targeting the EA community of practice, rather than the general user community which might benefit from the generic data infrastructure capacity. Another statement might be necessary for this other target audience (D4Science governance level).
Targeted messages to senior managers
develop the messages
Targeted messages to other stakeholders
information system specialists
Data managers face several challenges:
- Preferably work with only one DBMS (Database Management System) to reduce the time/cost of training and incresing efficiency;
- Use a reliable persistent repository, the responsibility of which potentially being outsourced, in order to focus only on the content of the datasets, not on the hardware/software issues; the repository must propose restricted access and confidentiality management, and ensure security against internal and external threats;
- Manage metadata at various levels, e.g., the ownership, contact, access rules, etc., or the general description of the contact. Both can change in time and the workload is not negligible when many datasets are assembled.
- Check/clean data and their internal consistency;
- Adapt the table/field structure for needs/requirements arising;
- Ensure inter-operabilty with between equivalent datasets and between datasets on various topics.
- All these operations are more difficult when the size of the dataset increase, in particular checking and cleaning, and important computing capacities are required.
Without entering in details at that stage, the D4Science/iMarine already or potentially answers all these needs. For example the infrastructure may propose a DBMS interface that will mask the original DBMS (MyQL, POstGres); the treatment of objects allows persistent metadata to be attached to the successive objects along a treatment; the links to other database e.g., between biodiversity and environmental parameters is ensured between VREs.
A case-show should be developed for advertising in the Board communities. It should entail an annual fee, details of it and costing to be worked out.
The capacity of the infrastructure to perform statistical analysis by accessing R libraries is very attractive in the case of highly demanding computing such as Markov chains, bootstrap, multivariate analyses with large datasets, etc.
Currently the development of the Bayesian approach in many, if not all, analytical scientific domains will lead to more demand on computing capacities.
The current show-case in the infrastructure is the computation of Length-weight relationships (LWR) for the 32,500 fish. These relationships are essential in fishery biology, but are not from original work for less that 15% of species (actually 3,588 species recorded in FB, and we know FishBase has a good coverage of the literature for that topic.
A second potential show-case is to apply these methods to many attributes/characters of all species, which wil require intensive computation.
Scientist have two needs with respect to infrastructure: - Assembling their own data and develop innovative tools of their own: they need to understand and control in detail the processes of the system, sometimes in very deep layers. - Using data and tools that they are not developing or experimenting themselves, often to nurture the ones they develop as above.
One example is the species distribution/niche modeling such as AquaMaps where the environmental parameters is usually not work out by the ecologist himself while he will put more efforts on occurrence data.
The infrastructure satisfies this dual need. However the first point requires collaboration with the technical team of the infrastructure, or to have staff knowledgeable in the development of new VREs/Portlets/Tools/Services/Parallelisation in the infrastructure.
While the second point may be already integrated in a business model, the second point requires more insights.
One showcase is the rapidity of answer by the infrastructure team to develop the environment for the Bayesian approach for LWRs.
Scientific results proposed directly to policy makers do not work usually, unless the latter have a string background in science.
There is a need of an intermediate level that digest the scientific results expressed in the scientific framework into something understandable by non-scientist.
The current showcase is a number of VREs developed by FAO for reporting and displaying graphically through GIS for instance the synthesis of results.
However, it is not clear if there can be a generic approach to develop VREs for science communicators. This has to be explored further.
If the business model integrates that level, and if there is no generic approach, the cost will be much higher because the technical group of the infrastructure needs to be involved each time.
- the Ecoknows type of story (from S.Garcia at AC2)
- through iMarine, “use our tools in the context of the developing world”, as well as “enhance our access to the knowledge in the developing world” (from J. Fischer at AC2) promote the case of the FishFinder standalone species fact sheet editor.
Inputs from iMarine May workshop to communication messages
The difficulty in understanding what the e-Infrastructure does is because it can do too many things. There is a need to segment markets and customers.
Difference between iMarine computing services, and MS Azure: The difference for the processing is that iMarine has specialized the way computing capacities are offered for the community that work in this environment. iMarine offers a tailored solution. And since we have worked both with Amazon and Azure in iMarine we have tried to make the adoption of these new technologies the easiest as possible. For example we have Hadoop clusters but the community doesn’t have to understand how to use Hadoop. This is why we are more community oriented and the other providers are too generic.
Imarine should be able to clearly demonstrate its functions and value it can add. There needs to be a stronger link between initiatives like imarine and policy development in DG Mare. What should be seen from iMarine is to take the current draft Data Collection Framework and state what can be delivered by iMarine, the associated costs of this and the quality of what can be delivered and the longevity of the services
Costs and benefits
A key aspect in defining and planning the sustainability of the programme is that of the cost and benefit characteristics, both in defining the cost contributions of various components of the programme and those for its ongoing development, and in assessing the costs and returns for potential end-users.
The first step of business model development is “what are the costs” (do not forget the “opportunity” costs of the providers of information) . Costing can focus on two aspects:
- minimum costs required to maintain operational the infrastructure: Important practical issues for sustainability concern the extent to which system elements can continue to be functional within evolving data platform developments, and if platforms are superseded, whether dependent applications can be cost-effectively migrated. The ownership, protection and conservation of data and meta-data sourced or applied by end-users also needs to be effectively addressed. Partnership arrangements will need to provide appropriate focus, adaptability, transparency, accountability and financial stability to ensure that long-term service outcomes can be successfully met.
- cost/benefit analysis of using D4Science at scenario/product level (e.g. Aquamaps)
A clear identification of benefits is requested; according to D4ScienceII sustainability workshop, key benefits are:
- lower costs / reduce the time to deliver
- offer new services/products that do not exist (a "Products catalog" will respond to the need to clearly inform communities on products and services as their packaging)
- capacity to better target users needs
- increase quantity and quality of data
- social benefits: connecting different research communities, supporting their collaboration and support interdisciplinary research.
Options should be explored for the establishment of the legal entity – country of establishment, form of entity, documentation required, costs of incorporation, duties and legal responsibilities of shareholders and appointed directors/officers, taxation and other issues, formal auditing and reporting requirements and the potential costs of meeting statutory needs. Issues concerning the legal/other acceptability and mechanisms for interacting with key partners and possible funders also need to be explored and confirmed. Based on the current membership and partner roles within D4Science and iMarine, France or Italy may be the preferred country of incorporation, but this would need to be reviewed