Traffic jams, road accidents and connected cows could all render existing network planning and optimisation tools unfit for purpose, if mobile operators don’t take action to make sure their systems are able to cope with the new demands placed on them by connected steers and steering wheels.
The issue is that current planning, monitoring and optimisation systems are designed to deal with devices and users with understood, predictable and often similar usage profiles.
Yet the Internet of Things and its cellular subset – M2M – introduces into that landscape a wider variety of device and usage profiles, and perhaps a greater chance of networks needing to deal with the unexpected, the unpredictable and the unknown.
Steve Bowker, VP Technology & Strategy, TEOCO, told TMN that the company had recently noticed the increasing profile of M2M devices in mobile networks in which its systems operate. Bowker said although overall volume is still low, the growth curve is exponential, and that led the company to think about how planning, monitoring and optimisation might adapt to the coming wave of M2M.
So the company commissioned Matt Hatton of IoT consultancy Machina Research to have a think about it – and Hatton’s take is that not only will cellular connections increase from 250 million to 2.3 billion in the next decade but that traffic will grow even more quickly from 200 petabytes in 2014 to 3.2 exabytes in 2024.
Importantly, although that’s only a small percentage of overall traffic volumes (4%) the nature of the traffic is different. Examples include connected cars that might bunch in large numbers at times of high traffic or during an accident – stressing a cell and core control elements with increased load as usage spikes, and potentially a signalling storm as devices all try and connect and reconnect to the network. The agricultural example – connected cows or moisture sensors to control irrigation – introduce a different profile into rural cells that until now have been sparsely used. And other use cases such as healthcare introduce a mission critical element to required service quality.
So are operators prepared for these new demands in terms of how they plan and optimise their networks?
Hatton: “Broadly speaking not really at the moment. They see it [M2M] as a nice extra add-on with no particular implications for the network, perhaps with the exception of some dedicated core network elements. It has not, in my experience, really entered the consideration of most network operators that they need to be thinking more carefully about the wider network impact.”
TEOCO’s Bowker said that what operators need to be doing is introducing analytics capabilities that allow them to identify exactly where trouble is occurring, but also the device profile of what might be causing the issue (and be affected by it), and the access channel involved. Added to that is an element of predictive analytics, so that in effect an operator is planning for and predicting an anomaly, rather than just reacting to it.
“Operators need a greater level of granularity,” Bowker said. “With geolocation to see where in the cell sites issues are occurring, and criteria so they can identify traffic types and by the nature of traffic. It’s the dynamic nature that is different, timescales are shorter compared to a long term view in which you might optimise by adding in a cell site or carrying out a cell site visit.”
Bowker added that it’s not a case of operators needing to spend a lot of money, but instead target where they are likely to be impacted, add analytics that let them know traffic characteristics and applying targeted design criteria into the design process.
Hatton concluded, “It’s about understanding the increasing diversity of devices you have to support, and the consequent characteristics and usage patterns.”
RESOURCES: “M2M growth necessitates a new approach to network planning and optimisation” – Machina Research’s paper on the planning and optimisation challenges of M2M.