In my last post, I covered cloud-native computing as one of several major trends in the Mobile industry I have been observing since joining Robin.io in May. Next, I’d like to dive into open architectures and how they are now critical to Mobile Network Operators (MNOs).
Trend #2 – Open Architectures
Beyond the cloud-native activities, every single mobile network function is being tasked to move to an open architecture model. CNCF, ONAP, OSM, ORAN Alliance, TIP, and others are all driving hard and fast towards open and interoperable standards. MNOs require flexibility to have multiple (even swappable) hardware and software vendors all while leveraging interoperable APIs to manage workflows and services across a both a flexible and abstracted architecture. I’m not sure 5G economics will work without it. If you think about the downward average revenue per user (ARPU) pressure MNOs have been battling, it will only continue with 5G. In fact, the pressure will increase given the level of investment, density and service innovation that comes with 5G. MNOs have to change the way they build and manage not just their networks but their service creation, provisioning and delivery. Migrating to an open architecture enables abstracted orchestration and workflow management.
Why does this matter? Vendor-specific scripts, protocols and interfaces ruin any chance MNOs have for full network scalability. Sure, one network application may be able to scale using the vendors’ proprietary methods. But if the OSS/BSS directs the network to instantiate a network slice, for example, the orchestration platform has to be able to identify available resources and deploy numerous applications, configurations and services in minutes, not days or weeks.
There is significant work being done on this topic in the area of Open RAN. I see that Open RAN is fundamentally about democratizing the RAN. Disaggregating the radio unit from the baseband processing functions (i.e. Distributed Unit or DU) as well as separating the real-time DU functions from the non-real time L2 and L3 functions such as Radio Intelligent Controller (RIC) and Central Unit functions, are fundamental to Open RAN. And running DU and CU software from different vendors on any properly specified commercial hardware is where significant technical and commercial benefits can be realized. I’ve also learned open doesn’t always mean OPEN. Some vendors are claiming to be open but are still slow to adopt full open and interoperable interfaces. That then hinders the benefit of the third trend I will cover in the next post – automation. In the meantime, please feel free to check out the Robin website for more information.