(Français) Projet STEEL

Secure and efficient data storage and processing on cloud-based infrastructures

— Preview

The STEEL project aims to provide solutions for efficient and secure data storage and processing on cloud-based infrastructures.

Gabriel Antoniu, Inria Research Director

Keywords : Cloud, data storage and processing, high-performance storage, confidential storage, hybrid cloud-edge infrastructures

Missions

Publications

Partners

Project website

The strong development of cloud computing since its emergence, and its massive adoption for storing unprecedented volumes of data in a growing number of fields, has brought to light some major technological challenges. In this project, we will address several of these challenges, organized into three axes:

The first involves exploiting emerging technologies for high-performance storage on cloud infrastructures, notably NVRAM-based, close to where data is produced and consumed (disaggregation principle).
The second area concerns the efficient storage and processing of data on hybrid, heterogeneous infrastructures within the digital edge-cloud-supercomputer continuum. This responds to the emergence of hybrid workflows combining simulations, analysis of sensor data flows and machine learning in many fields (autonomous cars, predictive maintenance, intelligent buildings, etc.). Their execution requires adequate management of storage resources, from the edge to cloud infrastructures, to enable them to be processed in a unified framework.
The third axis is dedicated to confidential storage, in line with the need to store and analyze large volumes of data of strategic interest or of a personal nature.

— Missions

— Our researches

STEEL is organized around three technical work packages. A fourth work package is dedicated to management, communication and dissemination of results.

Efficient, fault-tolerant storage

Take advantage of new storage technologies (NVRAM, CXL) to improve the performance and reliability of cloud storage solutions.

Storage on the edge-cloud continuum

Meet the data management and storage challenges of geographically distributed applications deployed across the digital continuum.
Propose a set of strategies to effectively deal with network partitioning, fault tolerance and extreme heterogeneity.

Confidential storage

The challenges addressed in this work package are: (1) storing ultra-sensitive data; (2) managing the heterogeneity of these secure enclaves; and (3) using them to handle massive data efficiently. These challenges will be addressed within the framework of this work package.

— Partners

Consortium

Inria, IMT, CNRS, Université de Bordeaux, Université Grenoble Alpes, Université de Rennes, INSA Rennes, INSA Lyon, IMT Atlantique

— Research team

16 permanent staff (not funded by the project)
17 people funded by the project, including:
8 PhD students
2 post-docs
6 engineers
1 contract researcher

Our teams in France

— Publications



        
            Contact

            
        
        
            Domains
        
            Keywords
Autoencoders Clustering Computing Continuum Computing Continuum Federated Computing continuum Data privacy Dynamic adaptation Edge Edge Computing Federated Learning Federated learning Framewoks High performance computing Key-Value-Store Key-Value-Store Trusted Execution Environment SGX Secret sharing Machine learning Memory Disaggregation Personalization Privacy SGX 
Federated Learning2
Federated learning2
Key-Value-Store2
Personalization2
SGX2
Autoencoders1
Clustering1
Computing Continuum1
Computing Continuum Federated1
Computing continuum1
Data privacy1
Dynamic adaptation1
Edge1
Edge Computing1
Framewoks1
High performance computing1
Key-Value-Store Trusted Execution Environment SGX Secret sharing1
Machine learning1
Memory Disaggregation1
Privacy1
Privacy and Security1
Reproducibility1
Robust federated learning1
Scheduling1
Secret sharing1
Stochastic model1
TEE1
Trusted Execution Environment1

        
            Authors
Costan Alexandru4
Antoniu Gabriel3
Prigent Cédric3
Cudennec Loïc2
Cudennec Loïc2
Simonet-Boulogne Anthony2
Ben Mokhtar Sonia2
Chelli Melvin1
Carastan-Santos Danilo1
Tedeschi Cédric1
Schubotz René1
Valli Mathis1
Ait Messaoud Aghiles1
Messaoud Aghiles Ait1
Boëzennec Robin1
Dufossé Fanny1
Pallez Guillaume1
Keahey Kate1

        
            Affiliated authors
        
            Journals
        
            Year of production
        
            Institutions
Université de Rennes5
Institut National des Sciences Appliquées - Rennes5
Université de Bretagne Sud5
École normale supérieure - Rennes5
Institut National de Recherche en Informatique et en Automatique5
CentraleSupélec5
IMT Atlantique5
Direction générale de l'Armement (DGA)4
iExec Blockchain Tech [Lyon]2
École Centrale de Lyon2
Université Claude Bernard Lyon 12
Institut National des Sciences Appliquées de Lyon2
Délégation Générale pour l'Armement2
Ministère des armées2
Université Lumière - Lyon 22
Institut polytechnique de Grenoble - Grenoble Institute of Technology1
Deutsches Forschungszentrum für Künstliche Intelligenz GmbH = German Research Center for Artificial Intelligence1
Argonne National Laboratory [Lemont]1

       
            Laboratories
       
            Departments
       
            Research team
    7 documents

Journal articles2 documents
Aghiles Ait Messaoud, Sonia Ben Mokhtar, Anthony Simonet-Boulogne. Tee-based key-value stores: a survey. The VLDB Journal, 2024, 34 (1), pp.10. ⟨10.1007/s00778-024-00877-6⟩. ⟨hal-04846840⟩
Cédric Prigent, Alexandru Costan, Gabriel Antoniu, Loïc Cudennec. Enabling Federated Learning across the Computing Continuum: Systems, Challenges and Future Directions. Future Generation Computer Systems, 2024, 160, pp.767-783. ⟨10.1016/j.future.2024.06.043⟩. ⟨hal-04659211⟩

Conference papers5 documents
Cédric Prigent, Kate Keahey, Alexandru Costan, Loïc Cudennec, Gabriel Antoniu. On the Reproducibility Challenges of Federated Learning: Investigating the Gap between Simulation, Emulation and Real-World Deployments. CCGrid 2025 - IEEE 25th International Symposium on Cluster, Cloud and Internet Computing, May 2025, Tromso, Norway. ⟨hal-04997547⟩
Aghiles Ait Messaoud, Sonia Ben Mokhtar, Anthony Simonet-Boulogne. TruShare: Confidential Key-Value Store for Untrusted Environments. 20th European Dependable Computing Conference (EDCC '25), Apr 2025, Lisbon, Portugal. ⟨hal-04935076⟩
Cédric Prigent, Melvin Chelli, Alexandru Costan, Loïc Cudennec, René Schubotz, et al.. Efficient Resource-Constrained Federated Learning Clustering with Local Data Compression on the Edge-to-Cloud Continuum. HiPC 2024 - 31st IEEE International Conference on High Performance Computing, Data, and Analytics, Dec 2024, Bengaluru (Bangalore), India. pp.1-11, ⟨10.1109/HiPC62374.2024.00033⟩. ⟨hal-04779813⟩
Robin Boëzennec, Danilo Carastan-Santos, Fanny Dufossé, Guillaume Pallez. Allocation Strategies for Disaggregated Memory in HPC Systems. HiPC 2024 - 31st IEEE International Conference on High Performance Computing, Data, and Analytics, Dec 2024, Bengalore, India. pp.1-11. ⟨hal-04815672⟩
Mathis Valli, Alexandru Costan, Cédric Tedeschi, Loïc Cudennec. Towards Efficient Learning on the Computing Continuum: Advancing Dynamic Adaptation of Federated Learning. FlexScience 2024 - 14th Workshop on AI and Scientific Computing at Scale using Flexible Computing Infrastructures, Jun 2024, Pisa, Italy. pp.42-49, ⟨10.1145/3659995.3660042⟩. ⟨hal-04698619v2⟩

Other projects

CARECloud

Understanding, improving and reducing the environmental impact of cloud computing

Read more

Archi-CESAM

Efficient and safe converged architecture based on near-memory gas pedals / Archi-CESAM

Read more

DiVa

Virtualized disaggregation

Read more

Taranis

Model, deploy, orchestrate and optimize cloud applications and infrastructures

Read more

PEPR Cloud projects

Spirec

Multi-level supervision and prediction for geo-distributed and heterogeneous infrastructures in the Cloud/Edge/IoT continuum

Read more

Steel

Efficient and secure data storage and processing on cloud-based infrastructures

Read more

Silecs

Super Infrastructure for Large-Scale Experimental Computer Science for Cloud/Edge/IoT

Read more

TrustinCloudS

Cloud infrastructure security

Read more

(Français) Projet STEEL

— Partners

— Publications

Contact

Domains

Keywords

Authors

Affiliated authors

Journals

Year of production

Institutions

Laboratories

Departments

Research team

Journal articles2 documents

Conference papers5 documents

CARECloud

Archi-CESAM

DiVa

Taranis

PEPR Cloud projects

Spirec

Steel

Silecs

TrustinCloudS

FORMULAIRE D'ABONNEMENT

Inscrivez-vous pour recevoir chaque trimestre la Newsletter du PEPR Cloud.

- Piloté par

- Financé par

- Opéré par

- Managed by

- Financed by

- Operated by