DESIRE6G will be present at the upcoming Optical Fiber Communication Conference (OFC 2024) from 24 to 28 March 2024 with two demos showcasing some of the latest developments in the project.

Demo 1: Deployment of Secure Machine Learning Pipelines for Near-Real-Time Control of 6G Network Services

In the rapidly evolving landscape of beyond 5G and 6G network services, Near-Real-Time (NRT) autonomous operation is crucial. Our solution focuses on decentralized control to minimize response times, deploying intelligent agents close to data plane devices while addressing security concerns through Distributed Ledger Technologies (DLT).

Key Highlights:
✨ Autonomous agents making near-real-time decisions based on distributed measurements.
✨ ML pipeline deployment and orchestration designed dynamically based on topological information.
✨ Integration of technologies ensuring the security of the ML pipeline, including the innovative use of DESIRE6G DLT network.

Demo Content & Implementation:
This demo showcases a complete control and orchestration system, capable of deploying secure ML pipelines and connecting distributed agents for the NRT control of highly dynamic NS. The main innovations of this demonstration are the use of autonomous agents making NRT decisions based on measurements collected and processed in different locations and supervised by a centralized entity, the deployment of ML pipelines that involves determining their optimal design based on topological data from a dedicated system after deployment, and the integration that provides the required security to the ML pipelines. A key aspect is the showcase of the use of the DESIRE6G Distributed Ledger Technologies (DLT) network, where new agents are dynamically associated to the DLT to exchange keys through a smart contract deployed for the NS.

This work is the result of the collaboration between partners Universitat Politècnica de Catalunya (Pol González Pacheco, Sima Barzegar, Davide Careglio, Marc Ruiz Ramírez, Luis Velasco), Universidad Carlos III de Madrid (Adam Zahir Rodríguez, Milan Groshev), and Telefónica Research Spain (Alejandro Muñiz Da Costa) with collaborators from University of Bologna (Chiara Grasseli, and Franco Callegati).

Demo 2: Distributed Multi-Agent System fed with Telemetry Data for Near-Real-Time Service Operation

In the realm of dynamic service operations, near-real-time capabilities are paramount. Our solution introduces a Distributed Multi-Agent System, empowered by telemetry data, to revolutionize service operations.

Key Highlights:
✨ Harnessing the power of a multi-agent system for near-real-time service operation.
✨ Telemetry data as a crucial feed, enabling intelligent decision-making across distributed agents.
✨ Addressing the challenges of dynamic service landscapes with our innovative approach.

Demo Content & Implementation:
This demonstration showcases the innovative capabilities and functionalities of P4-based switched networks. Specifically, the application of P4 In-band Network Telemetry (INT) telemetry to measure and aggregate per-flow and per-route QoS is a key innovation to enable multi-path routing, and the operation of a distributed Multi-Agent System. This architecture focuses specifically on NRT control as it greatly reduces response times, while liberating the centralized SDN controller from such tasks.
The systems in the demonstration are being developed within DESIRE6G, and will be deployed in a distributed federated testbed including the CNIT/SSSA (ARNO testbed) in Pisa (Italy) and the UPC testbed in Barcelona (Spain).

This work is the result of the collaboration between Universitat Politècnica de Catalunya (Pol González Pacheco, Sima Barzegar, Marc Ruiz Ramírez, Luis Velasco , Scuola Superiore Sant’Anna (Faris Alhamed), CNIT – Consorzio Nazionale Interuniversitario per le Telecomunicazioni (Francesco Paolucci), and NVIDIA Denmark (Juan Jose (JJ) Vegas Olmos).

We are working on some videos to share more details from these demos. Follow us on socials to stay up to date!