 |
|||
|
|||
|
|||
Home > TERATEC FORUM >Workshops
 |
TERATEC 2025 Forum |
|
Continuum HPC / Edge / Embedded
Chaired by Thierry Collette, Director Information Science and Technology Group, Thales and Eric Monchalin, Vice-President, Head of Machine Intelligence, Eviden
 |
At this workshop bringing together advanced industrial users, leading technology companies, and scientists, we will explore the challenges and opportunities presented by continuum computing. This groundbreaking approach aims to unify data processing capabilities, from traditional data centers or HPC (High-Performance Computing) to edge and embedded devices as well as IoTs (Internet of Things). We will address key issues related to resource management, latency, and security, while also integrating the latest advancements in AI for edge environments.Continuum computing aims to eliminate one of the primary obstacles to the full-scale digitization of businesses. As our digital world is continuously generating massive amounts of real-time data, the need for numerical simulations that can process these data streams is rapidly increasing. This demand calls for a seamless unification of computing resources to ensure efficiency and scalability.
Consequently, scientific and industrial infrastructures across various sectors are progressively shifting toward this approach. However, despite its clear advantages, continuum computing presents several significant challenges. Firstly, resource management stands out as a major concern. While HPC environments are typically designed to handle large, standalone, non-real-time workloads with centralized, unconstrained resources, edge devices often face significant limitations in processing and storage power. A critical challenge, therefore, is optimizing resource allocation and determining their appropriate sizing, taking into account both the data’s location and the needs of each use case.
Latency is another key challenge. Many critical applications, particularly those in IoT, require low-latency processing to operate effectively in distributed infrastructures. Developing new architectures and protocols to ensure fast and efficient processing of time-sensitive data throughout the entire computing chain is essential.
Security is also a crucial concern. The decentralization of data and processing across servers, edge devices, and IoT endpoints introduces new vulnerabilities and complicates the task of safeguarding these systems. The diversity of processing architectures adds layers of complexity to security measures and extends attack surfaces, making it vital to implement robust protocols that protect computations in progress, as well as data both in transit and at rest, while ensuring compliance with data protection regulations.
Finally, with growing prominence of AI, it is increasingly important to integrate AI application development and execution across both centralized and edge computing environments. This integration calls for the deployment of distributed “MLOPS” or even “MLSECOPS” pipelines, which were once confined to centralized systems, to ensure seamless operation and security in these more distributed settings.
Register now and get your badge here >>>
For any other information regarding the workshops, please contact :
Jean-Pascal JEGU
Tel : +33 (0)9 70 65 02 10
jean-pascal.jegu@teratec.fr
Campus TERATEC
2, rue de la Piquetterie
91680 BRUYERES-LE-CHATEL
France