This blog is by Volker Held, head of Industry Landscape at Nokia Solutions and Networks.
Operators spend about 10% of their revenue on operating, managing and optimizing networks. That’s already high enough, but networks are becoming even more complex and difficult to manage. New devices, applications and radio access technologies (RAT), not to mention small cells and diverse backhaul technologies, will potentially lead to higher network OPEX.
Constantly changing operating situations also make it difficult to predict network conditions. Meeting these challenges will need better knowledge of what the network is doing, with faster and better analysis and network optimization that is increasingly automated.
What if a mobile network could predict the movement patterns of commuters and allocate network resources automatically? Or identify an event with high data demand? The techniques of algorithmic network management will make this possible. Network elements will become aware of geographic activity. They will communicate their status to other network elements and will make vital operational changes on their own, without manual intervention.
Today, traffic steering is based on network element measurements like radio interface performance and service load. It is often done manually. In the future, traffic steering will be based on the statistics of customer experience, application and service history, real-time insight into current applications and subscriber segmentation, as well as predictive analysis. Future networks will automatically and quickly adapt to changing situations. They will become intelligent.
A cognitive network? Believe it!
Ultimately, evolution in data analytics, machine learning and machine reasoning will produce something new. Self Organized Networks (SON) and Customer Experience Management (CEM) will evolve into what we call a Self-Aware Network. SON and CEM solutions are already commercially available and support many analytics and automation use cases. Yet, much more can be achieved. New learning and reasoning algorithms hold out the promise of networks that can interpret highly variable and often incomplete data as well as monitor and propose actions.
A Self-Aware Network incorporating learning and reasoning algorithms will enable vast amounts of data to be handled and complex end-to-end network and service management tasks to be performed autonomously. This will help deliver the best customer experience at minimum cost, achieving the best business results for the operator.
NSN believes that using the full diversity of radio access layers and technologies in an orchestrated, customer-oriented approach is essential for building the cognitive networks of the future – networks that sense, learn, reason and act autonomously.
To learn more watch our webinar on: Technology Vision 2020 teaching networks to be self-aware.
We have more Innovative thinking to share at NSN here.