Sure, here’s the translation:
The OpenNebula platform is positioned as a key component for deploying a multicloud strategy of sovereign and distributed AI in Europe, combining privacy, performance, and regulatory compliance.
As Artificial Intelligence (AI) becomes the engine of digital transformation, the need to process data as close to its source as possible grows. Whether due to privacy concerns, regulatory demands, or just to improve performance, centralized approaches are giving way to a new architecture: decentralized AI. In this context, OpenNebula emerges as a strategic solution for deploying federated AI applications in multicloud and edge computing environments.
One of the most relevant use cases is FedX, a platform developed by E-Group that specializes in federated learning. This technology enables various institutions to train AI models without sharing sensitive data among them. The data remains localized in their isolated environments, while only model updates are exchanged, thus preserving privacy.
Virt8ra: Europe’s Decentralized Testing Ground
To demonstrate the viability of this proposal, OpenNebula has deployed FedX on Virt8ra, a pan-European infrastructure based on its technology. This testing ground is part of the IPCEI-CIS initiative, funded with a budget of 3 billion euros, and brings together cloud providers and research institutions from various European countries.
Each node within Virt8ra maintains its sovereignty over data and resources while collaborating in the training of multilingual AI models. The system has been tested with local data such as German legal texts, Spanish blog posts, and Polish newsletters, demonstrating its ability to build a global model without violating the confidentiality of the sources.
A Complete Technical Ecosystem for Federated AI
OpenNebula provides a set of key tools to enable these complex deployments:
- Ready-to-use orchestration, with Kubernetes through OneKE, its CNCF-certified distribution.
- Advanced networking capabilities, supporting public interfaces and VXLAN that facilitate secure communication between sites.
- Federated scalability, allowing the deployment of edge nodes in different regions, each with its own data and operational autonomy.
- Privacy-preserving training, where only model updates are transmitted, never raw data.
Additionally, the environment offers monitoring tools such as Prometheus and Grafana, distributed storage via Longhorn, and experimentation possibilities with JupyterLab.
A Commitment to European Digital Sovereignty
The project is supported by the Ministry for Digital Transformation and Public Function of the Government of Spain, as part of the ONEnextgen (UNICO IPCEI-2023-003) initiative, and is co-financed by the European Union through the NextGenerationEU funds.
This public-private collaboration not only strengthens Europe’s technological autonomy but also offers a practical and scalable solution for organizations operating in multiple jurisdictions and needing to comply with strict data protection regulations, such as the General Data Protection Regulation (GDPR).
A Decentralized Future for AI
With this advancement, OpenNebula solidifies its position as one of the most versatile platforms ready to tackle the challenges of distributed AI. The deployment of FedX on Virt8ra is not just a technological demonstration, but a roadmap towards a more respectful, sovereign, and resilient model of Artificial Intelligence, in which Europe aims to play a leading role.