Main features: inter-slice conflict and underutilization mitigation
Objectives: efficient resource exploitation in the RAN-Edge-Cloud continuum
Value proposition: efficient use of resource to maximise the return from leasing the infrastucture to multiple tenants
What is new: protocol learning as a basis
Addressing target audiences and expressing needs
- To raise awareness and possibly influence policy
We are looking for further adoption by the network automation community, and possible licensing of the developed solution.
- EU and Member State Policy-makers
- Research and Technology Organisations
- Academia/ Universities
R&D, Technology and Innovation aspects
Design of the AI virtualizer, Implementation of the AI virtualizer in simulation, Dockerization of the AI virtualizer, Integration with the slice manager, and Testbed validation
The developed AI virtualiser relies on the concept of multi-agent emergent communication, which is a generic framework that can be reused/replicated in various problems where several entities (e.g., multiple slices/vendors/operators) seek to negotiate/reconcile their conflicting objectives.
The result is sustainable by design as it enables the maximisation of the use of the infrastructure, reducing OPEX.
