Share

A device-to-cloud continuum for IT-OT convergence

Credit : CEA
One of the main obstacles to the massive end-to-end digitalization of business processes is the continuity of data from the field to the cloud. The purpose of the OTPaaS project is to develop a high-performance operational data management and storage platform for industry that is capable of delivering the required response times, security, trust, and energy efficiency.

The project, which kicked off in 2021 for an initial three-year period, is being run by a consortium of leading national industrial corporations (Atos, Schneider Electric, and Valeo) and small to medium-sized businesses (Agileo Automation, Mydatamodels, Dupliprint, Solem, Prosyst, and Soben). Top-tier research organizations (CEA, Inria, IMT, and CAP’TRONIC) are providing substantial support. CEA-List, CEA-Leti, and CEA-Liten are contributing to the project, which is part of the French government’s national cloud computing strategy and the national Industry 4.0 roadmap. As project coordinator, CEA-List is getting all its departments involved in developing solutions to manage the data produced by cyber-physical systems.

During the first year of the project, the objectives were confirmed, and the research began, with CEA-List focusing on three main pillars:

1- Manage heterogeneous distributed infrastructures from device to cloud (Figure 1):

  • Identify major classes of industrial IoT architectures and associated hardware and software.
  • Develop a methodology for navigating between these architectures according to system-level constraints.
  • Test Kubernetes for building a container orchestration platform from Edge (OT) to cloud (IT).

 

2- Develop technology bricks for the Edge-to-cloud continuum.

3- Improve and optimize the use of our technology bricks in the trusted IOT2cloud continuum:

  • Our technology bricks will be integrated into the Edge-to-cloud continuum for an industrial waste management use case (Figure 2).
  • Substantial research went into selecting the technology bricks, with ROS as the middleware for integrating robotics.

 

Applications will be developed as ROS nodes and deployed as containers. A variety of protocols will then be used for operational deployment and maintenance.

 

Figure 1: Generic Blueprint

 

 

Figure 2: Waste sorting management architecture

 

We are making distributed systems more reliable from Edge to cloud.

Rebecca Cabean

Fabien Baligand

Research engineer — CEA

The technology in use:

  • A model for the waste management demonstrator was completed in Q3 2023. An initial demonstration will be carried out at DupliPrint in 2024.
  • We are also consolidating the R&D work initially completed for OTPaaS for the French government’s PEPR Cloud initiative. The groundwork is being laid for the next phase of development with partners Valeo, Schneider Electric, and Prosyst through the Data4Industry-X project coordinated by DAWEX.

Major partnerships:

  • Atos/Bull, Schneider Electric, Valeo, Agileo Automation, Mydatamodels, Dupliprint, Solem, Tridimeo, Prosyst, Soben, Inria,  IMT, CAP’TRONIC.

See also

Software development environments

N2D2

N2D2 is used to optimize neural networks, embed them on components or on dedicated hardware accelerators, and measure their performance.
Read more