At present the O-RAN architecture provides a promising solution of an open-RAN ecosystem, where based on the defined functional splits (CU, DU, RU) a multi-vendor solution can theoretically be achieved. This so called “wave 1.0” 5G that is capable of utilizing only basic (rough) virtualization as well as introducing essential interfaces to enable open-ecosystem, like: E2 for the control of CU/DU/RU as well as A1, O1, O2 for policy based management, network configuration and monitoring. The existing state-of-the art based on IS-Wireless analysis and experiences (also as O-RAN member) should be upgraded to what we call open-RAN Wave 2.0 in order to allow greater flexibility of functional split as well as improve the capability of addressing the challenges of ultra-dense networks. Flexibility of functional splits is essential to adjust open-RAN based networks to the existing infrastructure capabilities including not only fronthaul but also midhaul interfaces. Fronthaul is understood mainly as splits beyond 6 and especially the 7.2 O-RAN split that requires a certain level of capacity, which may be even quadrupled with the split 7.1. In the midhaul e.g. where the CU-CP with RIC (RAN intelligent controller), CORE, MEC and application servers are located, the infrastructure can also vary in capacity. With highly granularized network functions packaged as VNF/CNF (virtual machines of containers) and also providing multitude of split options it is easier to tailor deployment of open-RAN network to fit into available fronthaul and to optimize cost of hardware and network. Moreover, it is then more convenient to orchestrate such “workloads” (i.e. 5G radio stack functions) across edge-cloud continuum, also including edge micro data centers. In this way, multiple split association types can also be achieved naturally e.g. split per slice, per UE, per bearer. The underlying compute resources can also be utilized more efficiently as particular workloads can be fitted to a variety of acceleration cards (GPU, FPGA, SmartNIC) or computer architectures (x86, ARM). Eventually such fine grained, highly composable (orchestrated) disaggregated open-RAN can be called open-RAN Wave 2.0, as it enables achieving higher capacities for network operators who are aiming to address the challenges of ultra-dense networks. Efficient data-driven resource management (both radio and compute) with the novel paradigms like cell-free (or distributed cell-free massive MIMO) are becoming more straightforward to be implemented with such improved open-RAN architectures.