Future RAN evolution: Tools to support strategic decision making

Radio Access Network (RAN) evolution is as important as ever.

Currently, Mobile Network Operators (MNOs) operate networks that comprise of 2G, 3G and 4G technologies, while there are extensive preparations for 5G (see for instance European Horizon 2020 projects FANTASTIC-5G, Flex5Gware & Speed5G among others).

As requirements and complexity rise, MNOs need to take timely and well informed strategic decisions about the future usage of spectrum and technologies in their network. Questions like the one posted in https://www.telecomtvtracker.com/insights/2g-3g-retirement-what-is-the-correct-timing-8187/ are becoming more important than ever.

“What is the correct time for the decommissioning of 2G and/or 3G technologies?”,

“When should a mobile network operator start to re-farm the respective 2G and/or 3G bands to the 4G network, so as to increase profitability?”,

“What will the traffic, capacity and spectrum requirements be in the next 2, 5, 10 years from now?”, “Is there a first mover’s advantage here? “

MNOs taking critical decisions on spectrum too early, may end up losing important customers to competitors and generating customer frustration. However, taking actions too late equals to lower operational efficiency, hence disadvantage to competition that can utilize the additional capital to invest more aggressively than the late mover. And there is more.

- Are all mobile network markets the same?

- Are all locations of a market similar?

- Is the optimal time common among all cells of a specific network?

- How does the individual strategic positioning of each MNO influences such decisions in competitive telecom markets?

- Should this be a waterfall approach or could it be an ongoing process, or even dynamic allocation?

The answers to these questions are hidden under countless layers of complex techno-commercial factors and details. Briefly speaking, they are hidden under vast amounts of highly heterogeneous data (“Big Data”). Customer behaviors over time and space, voice and data usage, application usage profiles, mobile device availability, network element growth, supporting technologies along with the predicted impact in the network KPI / KQIs are only some of the factors that can reveal the impact of alternative decisions in short-, mid- and long term.

Incelligent has proved that this demanding analysis can greatly benefit from its unique deep learning – driven predictive technology. The approach is shown in the figures below:

Ran Decision Support Charts by Incelligent

The Incelligent predictive models, by optimally exploiting the right heterogeneous data stemming from market & network sources, support a complex ongoing process that makes possible the alignment between the strategic corporate goals and the selection of spectrum bands to be used in the journey from 2G to 4G networks and beyond.

Fast and accurate predictions, through patented proprietary algorithms, are achieved (ranging from 85-95%), leading to higher (by at least 25%) revenues.

The charts above come from a real - life use case of a market - leading mobile network. 

Overall, Incelligent develops big-data analytics and predictive tools for supporting strategic decisions including future evolution of RAN.