AI is helping to reduce the cost of providing and maintaining broadband for heavy traffic areas
At the Ultra-Broadband Forum hosted by Huawei in Geneva Switzerland Heinz Herren, the CTIO of Swisscom informed the audience that they will be bringing 1 gigabytes of fixed connection to all of Switzerland.
The plan is to have the infrastructure to achieve this in place by 2021. Switzerland has a population of over eight million, many of which live in rural isolated areas, Zurich the countries largest city has a population of 425,000
Bringing high speed connection to mountainous areas will be no easy task. Swisscom in a partnership with Huawei are currently putting the infrastructure in place and in urban areas have set up fiber optical cables which run directly into people’s business or homes, often referred to as FTTH connections.
The rest of the connections to rural areas will be done via FTTC connections which involve utilising the existing copper networks in place to cover the distance from street to house. This will be done using DSL technology G.fast.
G.fast is the successor of VDSL2 and the technology enables speeds of up to 1 Gigabyte However, G.fast only works efficiently over 250 meters.
In his keynote speech at the conference Ryan Ding Executive Director and CEO of the Huawei Carrier BG emphasised the need to build stronger broadband infrastructure to deal with industrial and commercial requirements.
He commented that: “This year’s World Cup was broadcast in 4K, and the 2020 Olympics in Tokyo will be transmitted in 8K. That demands 100 mega bandwidth.”
“Smart traffic video cameras used in urban transportation need higher uplink bandwidth. Cameras across three lanes require about 100 mega upstream, and the full monitoring of a crossroads needs around 500 mega.”
For broadband providers he highlighted how important machine learning and artificial intelligence innovation are proving to be when it comes to system efficiency and cost reduction.
Mr Ding stated that AI has helped reduce alarms and home vists through predictive analytics of home networks and cut trouble shooting by 30 percent.
He also noted that: “By analyzing and predicting the relationships between traffic, power consumption, and cooling, AI O&M (operation & maintenance) can reduce power consumption by 10 to 20%, without affecting network KPIs or KQIs.”
Lastly he told the audience that Huawei’s work using AI has seen data flow efficiency increases of 30 percent. This is achieved when AI is used to analyse traffic bottleneck.
“AI will significantly improve O&M efficiency and the user experience. For network faults that are hard to replicate, we can use big data to replay them and locate the root cause. This means we no longer need to go onsite to diagnose the fault,” Mr Ding added.