HexRIC: Building a Better near-Real Time Network Controller for the Open RAN Ecosystem.

The Open Radio Access Network (O-RAN) architecture has served as a key enabler for the programmable control of cellular networks through its introduction of the RAN Intelligent Controller (RIC). However, the RIC solutions available today are hobbled by scalability issues, inaccessibility, and a general lack of robustness, thereby serving as an impediment to the widespread adoption of O-RAN. To that end, this paper introduces HexRIC- a purpose-built next-generation network controller for the O-RAN ecosystem. Key highlights of HexRIC include a robust messaging infrastructure, a machine learning operations framework, and support for O-RAN's flagship use cases such as RAN slicing. Furthermore, the paper also includes a comprehensive system evaluation addressing the key themes of scalability, ML model lifecycle management, and finegrained network control. The results thus obtained showcase that not only is HexRIC the most feature-complete control platform to date, it also offers a significant performance advantage over the prior art.