Intelligence for next generation networks and societal challenges
The level of integration of mobile networks in our day-to-day activities (e.g., automotive, extended reality (xR), smart home) is making them become the nervous system of our society. Hence, 6G networks will become a complex system for which manual management as traditionally done is not an option. The enabling technologies and intelligence used to handle such complex system may as well be of application to tackle complex societal challenges (e.g., eHealth, smart cities). Our group is conceiving such a societal nervous system.
Intelligence to build the future service-network-cloud continuum
The goal of SaS is to design the architecture and algorithms of xG networks serving demanding networking use cases, such as automotive, extended reality (xR), (industrial) IoT, and those tackling societal challenges (e.g., eHealth).
Its main objectives are:
- To contribute to the digital transformation of society by designing the networks (incl. all segments, such as cloud, edge, fog, transport, mobile core, mobile RAN) that future vertical applications need. In this way, a services-network-compute continuum will be built.
- To make an efficient use of end-to-end network resources (network, computing, storage) by enabling automatic and autonomous management and orchestration mechanisms and secure inter-relation with involved administrative domains.
- To exploit AIML techniques to extract knowledge from all sorts of data sources and integrating it into network and/or vertical application workflows, hence generating added value going beyond the scope where the data was generated
- To evaluate the application of the techniques used in research fields beyond networking that show potential (e.g., vehicular, IoT, eHealth, smart cities).
- To assess its research findings in close-to-real experimental scenarios, which includes the construction of the experimental research frameworks that are required (incl. EXTREME Testbed® and PARADIGMS Testbed) and their integration with those of other groups.
Research lines
Centralized and decentralized management and orchestration (MANO) architectures to support automation of complex, heterogeneous, distributed, shared infrastructures (including open RANs) over which service meshes are deployed. Network automation is based on the integration of closed-loop decisions involving monitoring, data engineering, and AIML (e.g., for SLA management).
Furthermore, the entities and interfaces devised support end-to-end slicing over multiple administrative domains involving an increasing number of stakeholders building trustful administrative relationships through distributed ledger technologies (DLT). Proofs-of-concept are developed over the EXTREME Testbed® and other experimental frameworks of the group (e.g., Paradigms) featuring heterogeneous computing and networking infrastructures (incl. a 5G mobile network) to emulate a variety of industry-oriented scenarios.
Algorithmic framework of beyond 5G/6G networks devised in the previous line targeting cognitive management engine, decentralized management/decision-making, service optimization, information abstraction, control-loop management, SLA management, intent-based networking, anomaly detection and restoration, or slice isolation.
This line exploits the latest advances in artificial intelligence/machine learning (AIML), such as federated learning, graph neural networks, transferrable AI, and exploits this expertise to also target problems beyond the networking field (e.g., eHealth, smart cities) towards the practical realization of artificial general intelligence.
