“…Moreover, they have argued that recent advances in ML enable building some of NDT's core components as a data-driven network models that can operate in a real time, including a routing optimization in a QoS-aware use case. The other researchers reported on a novel framework of digital twin manager dedicated to handle conflicting network applications [31], and also designed a digital twin suitable for "what-if" analysis in Border Gateway Protocol (BGP) optimization [32]. Finally, the paper [33] commenting on 5G/6G network softwarization and intelligentization, outline the role of digital twin architecture for network autonomy, predict an emergence of a service layer in 6G networks compatible with digital twin and able to realize proactive analytics, including generative intelligence functionality.…”