Looking into the future of telecommunications, it could be argued that AI and telcos will effectively transform each other, explains Raf Meersman
The year 2020 has been one of countless changes. Changes that have been happening in the past 10 years have taken a quantum leap in the past 10 months.
Can you believe that last year, many of us flew halfway around the world for a one-day meeting? That remote education, remote first line health advice, or even working from home (or anywhere) seemed like a distant future?
The world has adapted rapidly. Operators worldwide reported overnight increases in bandwidth usage between 20 and 50 per cent. It is our sector that was able to provide the infrastructure, services, flexibility to support this change without any major hiccups.
Look in the mirror
The telecom sector was ready for this sudden ‘fast forward’. Just like artificial intelligence (AI) is ready for its breakthrough. Companies have been adopting AI faster, to minimise the spread of the coronavirus by human contact.
Our industry is the catalyst of this adoption. But have we looked in the mirror? Are we aware of how these AI innovations affect our own industry? If we are willing to embrace the benefits of digitisation through AI, I see an industry that can become an innovation leader.
Today, there is already a myriad of AI applications available in all stages of the network lifetime.
If you remember, even back in 2010, Swisscom claimed that its robot driven fibre grid deployments yielded a 50 per cent cost saving. If cars can drive autonomously, why can’t digging machines do the same?
Moreover, on the monitoring side, there are AI techniques being developed which can predict faults and even make equipment self-heal, like the automatic restart of a server. AI capabilities can also be used to enable network function virtualisation (NFV). NFV allows infrastructure owners to offer access to its equipment as a service to altnets. This means the birth of infrastructure as a service (IaaS).
Let us not forget about AI possibilities to protect networks. Self-learning algorithms can discover how to better detect and block intrusions on devices embedded in the network. As 5G introduces countless new antennas and connected devices, it will become much more vulnerable to attacks.
Practical Optimisation is key
There is also the planning and design stage of network deployments. This is a subject I care a great deal about, and I am sure we will see radical changes here as well.
We already see AI-based innovation in the gathering and cleaning of digital data where field surveys, for example, could be replaced by virtual reality sessions.Walking the streets from behind your desk, combined with automated recognition of sidewalk borders, survey types, poles, etc. Intrigued yet?
We have been on the AI track for more than 15 years now. As a next step we are adding AI technologies based on machine learning to the software. Practical optimisation is key here: optimising cabinet and closure coverage and sizes; optimising cable and duct routes; optimizing demand point connections by finding the right cost-coverage threshold per connection, etc... All this can be underpinned by smart algorithms that understand more than just the logic of mathematical formulas. They need to learn from the best network designers about the balance between cost optimisation and constructability. Here lies our investment, our desire to improve.
The role of the telecom industry has grown significantly this year. And it has the potential to continue to do so. Let's implement new technologies and inspire the rest of the world. How is your organisation contributing to the future of the telecom industry?
Raf Meersman is CEO at Comsof
