EOSC Future engaged with the private sector in order to develop suitable models for delivering commercial cloud services to Europe’s research communities through collective framework agreements and other procurement mechanisms. It also expanded the EOSC Digital Innovation Hub (EOSC DIH), which was set up in the previous EOSC-Hub project, as the primary platform for industry usage of EOSC resources. Throughout EOSC Future, efforts were also made to raise awareness of EOSC among innovative private enterprises through workshops, events and sharing success stories.
EOSC Future procurement activities
In order to stimulate digital service uptake, EOSC Future allotted a total of €4 million for the distribution of commercial cloud services via EOSC. This adoption funding was spread across procurement calls, aiming to:
- make agreements with commercial services available to all European researchers, at scale, in a reasonable amount of time
- ensure funding reaches across different domains and scientific communities
- bring advancements and innovation to diverse areas of research
- boost supplier engagement enable cloud access to valuable resources as Quantum computing, HPC, and Virtual Research Environments
A critical component was the introduction of ‘digital aggregators’. These entities could be research institutes, e-research infrastructures, NRENs or other research support bodies, already involved with the European research community. Since these entities already understood the needs of this community, their dedication was key to ensuring the integration and uptake of contracted commercial services from the selected providers.
Additionally, a dedicated EOSC Future procurement call focused on the demonstration of mechanisms by which procured public sector/commercial data resources could be registered as such through EOSC.
What did we learn?
- European RIs will continue to consume commercial cloud services to underpin and enhance their own services in order to increased scalability and fault-tolerance, for increased quality of service and to shape future sustainability
- There is a clear demand for commercial cloud services within the EOSC community
- There is a clear willingness from cloud service providers to partner and collaborate with digital aggregators
- There is a proven value of researcher/industry partnerships in driving innovative research solutions (i.e. development of RI services)
- A collective portfolio of agreements with commercial services for thousands of public-sector research-performing organisations requires scalable processes for demand assessment, joint procurement and agreement delivery.
- Digital aggregators can be valuable distribution channels to reach researchers at scale and unburden them from the financial-administrative aspects of using framework agreements
Read more about the lessons learnt in commercial cloud procurement from the OCRE project to EOSC Future (also included in Issue 45 2024 of the GÉANT community magazine).
Results of the EOSC Future procurement scheme
CALL #1: EUROPEAN RESEARCH E-INFRASTRUCTURES (ROUND 1)
5 grants, each of up to €400 000, were awarded to digital aggregators collaborating with OCRE cloud service providers to craft new approaches for the distribution of commercial cloud services available via EOSC.
1. OpenScienceLabs for HPC (Austria), from Sparkle and the Vienna Scientific Cluster
The HPC Development-OpenScienceLab is an online tool that senior researchers can use to simulate and create a HPC (high-performance computing) workflow, a feat usually out of their reach, due to limited computing power and resources. It is easy to use and includes sample data. Researchers can also use the tool to create HPC Analysis-OpenScienceLabs, which can be used to review scientific papers interactively, with real data and the researcher’s own analysis tools. Every OpenScienceLab runs on a specific Google Project that is the smallest logical “container” of cloud resources for Google. The proposed solution offers a user-friendly and secure environment for researchers to carry out HPC workflows and share their scientific findings in a transparent and reproducible manner.
2. Hate Classifier service (UK), by Quistor and the University of Sussex Applied Language Modelling Facility
This tool helps researchers who are working to analyse and classify hate speech – a difficult field to study because of the lack of clear definition of what hate speech is, creating unique challenges and specific needs for researchers. Hosted in a cloud-based environment and using high-performance OCI Nvidia GPUs for machine learning, the Hate Classifier service can be used by researchers to develop and train new models that could change the way we understand hate speech, setting a new standard for the use of machine learning.
3. European Environment for Scientific Software Installation (Netherlands), by Bechtle, Quistor as subcontractor and SURF
Based on the principle of infrastructure-as-code, this installation will help researchers to easily install and use scientific software by providing easy-to-read scripts that can generate the necessary research environment. These scripts can be updated over time and researchers can use different types of scripts to distribute their software more easily. Using this tool will make it easier for researchers to access public cloud infrastructure options while maintaining security through the SURF Research Access Management system.
4. INCD Cloud services (Portugal), from Sparkle and the Portuguese National Distributed Computing Infrastructure
This platform provides a single entry point for researchers to easily deploy and manage computationally demanding applications by allowing users to containerise workloads and create computing clusters that work across multiple clouds. GCP’s EC3 facilitates the multi-cloud deployments providing integration between the Google platform and the INCD infrastructure.
5. GWDG TRE (Germany), by Rackspace and the Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen
TRE is an open-source tool that helps research organisations create secure environments on the AWS Cloud for collaborating scientists from multiple research centres to perform their analysis, also providing a governance process to help researchers meet their compliance requirements across multiple jurisdictions. GWDG TRE aims to solve current issues that hinder the research community’s access to cloud-based agility, scalability, and cost efficiency by providing a common EOSC integrated infrastructure based on the Aggregator’s Landing Zone and the AWS TRE platform.
CALL #2: EUROPEAN DATA SPACES
For the second procurement call, EOSC Future welcomed innovative ideas on how to incorporate commercial and/or public sector data into thematic and discipline-specific data spaces in the EOSC context. This could involve making data available through specific methods, by virtue of the metadata, or incorporating data from various sources for a research use case.
1. iCRAG – SFI Research Centre in Applied GeoSciences:a high-resolution radiometric data from around sources of Irish building materials (Soil geochemical database)
2. PSNC – Poznan Poznan Supercomputing and Networking Center /AGRI Dataspace: Synthetic data (RGB imaging) of the diseases related to beet cultivation and wheat crops – GreenMatterAI
3. Chalmers Uni of Technology (Sweden)/ Eastern Switzerland Uni of Applied Science: Wind turbine operational and structural performance data
4. GWDG – Gesellschaft für wissenschaftliche Datenverarbeitung mbh Göttingen: Research data in the humanities is from the beginning of the disciplinary field created and discussed in books, journals, and other publications. Books and other printed documents are a big part of our cultural heritage. Though the literary system and the book market is regulated by licences and copyright which did not change during the transition into the digital age.
CALL #3: EUROPEAN RESEARCH E-INFRASTRUCTURES (ROUND 2)
Given the interest around Call #1, the project decided to focus its third instalment on partnerships between research infrastructures (digital aggregators) and commercial cloud providers (OCRE framework contract holders). From 12 exciting proposals, 2 were awarded €600,000 each (VAT excluded).
1. D4science – EU Virtual Research Environments On Cloud (Italy), from TI-Sparkle and Consiglio Nazionale Delle Ricerche (CNR)
The National Research Council (CNR) is the largest public research institution in Italy, the only one under the Research Ministry performing multidisciplinary activities. D4Science aggregates “science domain agnostic” service providers, as well as science community-specific ones, to build a unifying space where the aggregated resources can be exploited via Virtual Research Environments and their services. These include core services supporting VRE management, resource management, authentication, and authorisation; Data space management services; Data analytics services; and Collaborative services.
The major advantage for researchers of working with a VRE is that they find in the front-end a pre-configured environment with the research tools they need for their research (i.e. RStudio) and these tools run on IT resources (VMs, CPUs), scientific software packages and settings allocated by D4Science to the specific VRE according to the needs of the researcher’s group/community.
2. Quantum Compute Services On Cloud (Ireland), by Rackspace and the Munster Technological University
Quantum computing is a disruptive technology that harnesses the laws of quantum mechanics to solve problems outside the scope of our current classical computers. Quantum computers will be a core component of future networks, offering capabilities in cryptography, networking, data analysis and optimisation.
AWS provides access to multiple varieties of quantum computing hardware including the Gate-based superconducting processors, Gate-based ion-trap processors and Neutral atom-based quantum processors. Researchers can run experiments on a variety of quantum hardware, comparing and contrasting their results. This would not be possible without cloud-based quantum computing access, which significantly strengthens the results of experiments and proof of concepts.