Design of AI-based Digital Twin Network for Multimedia Service Provisioning

In the era of pervasive connectivity, the effective provisioning of multimedia services surely represents a cornerstone challenge for both infrastructure and service providers. Given the recent evolution of Beyond 5G/6G networks, it is conventionally accepted that a Network Operator (infrastructure provider) shares its resources with tenants (service providers) by giving them the possibility to autonomously and efficiently configure the application services they provide to end users. To this end, this contribution proposes a cutting-edge framework for multimedia service provisioning, that harnesses the power of Artificial Intelligence (AI) in conjunction with Digital Twin Networks (DTNs). Specifically, it considers the real network features collected from the physical real network through Software-Defined Networking Controllers to build the DTN, whose purpose is to provide present and future physical, network, and application statistics through AI for efficient and proactive resource allocation. The DTN exposes to the tenant a concise set of monitoring parameters related to actual network-specific statistics and Quality of Service metrics and also future ones thanks to Deep Learning, without sharing in-depth details of the underlying network through Machine Learning (clustering). This information will be exploited by the service orchestrator for optimal resource allocation and redistribution, which will be deeply investigated in future works.