BDVA workshop on new data architectures

BDVA hosted a dedicated workshop to review the evolving landscape of data architectures. The event explored emerging trends and examined how new approaches address upcoming challenges.

Share this news

BDVA hosted a dedicated workshop to review the evolving landscape of data architectures. The event explored emerging trends and examined how new approaches address upcoming challenges, including the technical and organisational complexities of data sharing and the implications of AI integration. The event gathered 98 participants.

The design of data architectures has undergone a remarkable transformation in recent years. What was once a relatively standardised ETL (extract, transform, load) process linked to a central data warehouse has evolved into a diverse ecosystem of complex frameworks, including data lakehouses, data meshes, data fabrics, data virtualisation and cloud-native models, often deployed in combination.

These modern approaches have emerged to overcome the limitations of traditional systems, particularly when dealing with the challenges of today’s data landscape: vast data volumes, unstructured and multimodal formats, the growing need for real-time processing and demands for scalability and flexibility.

Beyond these technical evolutions, the increasing need to share data across organisations, sectors and borders – to fully unlock its value – has spurred the rise of data spaces, which modern architectures must now be designed to support and connect with. At the same time, the primary application of data is shifting from traditional analytics towards advanced artificial intelligence (AI) workflows, bringing a new set of requirements for data management and infrastructure design.

The BDVA workshop “Evolving data architectures for distributed, intensive, AI-driven applications” (as part of AG71)

Date: 22 October, from 10:00 – 13:00 CEST 

Full agenda :

Time Topic Speaker Organisation
10:00 10:10 Introduction and BDVA news Ana García / Sinna Rissanen / Daniel Alonso BDVA
 

Distributed architectures for data sharing / data spaces

 

10:10 10:20  

Inspirational talk: integration of data spaces and agentic AI

Joaquin Salvachúa UPM
10:20 10:30  

Seamless cross-federation data access and management using Onedata

Lukasz Dutka Cyfronet
10:30 10:40  

Edinburgh International Data Facility (EIDF)

Jano van Hemert UK National Supercomputer Centre, EPCC, University of Edinburgh
10:40 10:50  

TEADAL: an architecture for trustworthy, secure, and federated data meshes

Pierluigi Plebani Politecnico de Milano
10:50 11:00  

ENVRI-HUB NEXT – Federating environmental Research Infrastructures across Europe

Marta Gutierrez David EGI Foundation
11:00 11:30  

Discussion

 

Architectures for AI

 

11:30 11:40  

Setting up the scene: AI Architectures for the Next-Gen Big Data Stacks

Raul Gracia Dell
11:40 11:50  

How data spaces fit in the data room: the case for Green.Dat.AI

Arturo Medela Eviden
11:50 12:00  

Abstraction of computing continuum to facilitate the deployment of AI workflows, especially when heterogeneous architectures are involved

Eduardo Quiñones Barcelona Supercomputing Center
12:00 12:10  

Serverless Data Processing of massive unstructured data for AI workflows

Pedro Garcia Lopez Universitat Rovira I Virgili
12:10 12:30 Discussion
12:30 Closing

Objectives of the Session

  • Review the current landscape of big data architectures, from established models to cutting-edge approaches.
  • Identify and analyse new trends shaping the design of modern data ecosystems.
  • Discuss challenges and requirements these architectures must meet, particularly in relation to:
    • Managing massive, unstructured and multimodal data.

    • Supporting real-time and AI-driven workflows.

    • Enabling cross-organisation and cross-sector data sharing through data spaces.

    • Addressing broader issues such as interoperability, governance, standards, security and trust.

  • Exchange insights and experiences among BDVA members to better understand adoption pathways, practical barriers and opportunities for collaboration.

This workshop fostered dialogue and collaboration among experts and stakeholders to ensure that Europe’s data architectures remain at the forefront of innovation, supporting trustworthy, interoperable and AI-ready data ecosystems.