Hey, what’s that IoT device on my network?
For years, we’ve talked and written about the Internet of Things (IoT) in the future tense. Not anymore! It’s happening now and growing rapidly. Today, there are already about 70 million cellular IoT devices around the world with an average annual growth rate of 14.3%. 1
To understand the effect these devices might have on the network, consider the amount of signaling typical of an IoT device compared to a smartphone. On average, IoT devices use three to ten times the amount of signaling that an Android device uses. They have the potential to storm the network with signaling and drain radio resources to the point that the service quality of non-IoT subscribers is affected. Given that by 2020 they will vastly outnumber consumer mobile devices, for the unprepared, IoT could fast become the Internet of Disconnected Things.
Is device detection in order?
To understand the networking resources consumed by IoT and the potential optimization needed, an operator needs to detect all the IoT devices on its network. IoT device detection is much easier when the operator is managing those devices and aware of each device type, but this isn’t always the case. Many IoT devices are unmanaged, which means that they’re not provided by the operator and only require an internet connection. Such unmanaged devices can create many challenges for the operator with regards to signaling and radio resource consumption.
Classic analytics solutions cannot identify the application used by the device or even the exact device type (e.g., drone of vendor X, camera, baby monitor). They simply report it as a generic mobile device (e.g., Sierra Airlink Gx440). Fortunately, the Nokia Cognitive Analytics for Mobile Networks can identify and classify unmanaged IoT devices based on their behavior and the behavior of the peer internet host in the network; this is done using a complex algorithm developed by Bell Labs.
Different IoT devices consume different network resources
Some operators are building overlay core networks to carry the traffic of IoT devices managed by the operator’s device management solution. Unfortunately, this is not sufficient because:
- Radio resources are shared between IoT and non-IoT devices
- Some subscribers have unmanaged IoT devices.
Each IoT device or application has a different impact on the network that depends on the function it provides. For example, a video surveillance camera normally requires an always-active session and high bandwidth to live stream good quality video without interruption. It doesn’t consume lots of signaling resources but heavily consumes radio resources. On the other hand, an electricity meter might need a short, low-speed connection every hour, which makes it chattier and consumes more signaling resources than the surveillance camera.
It is crucial in network planning to monitor the resources consumed based on the growth rate of each device type or when a new device type is introduced. The operator needs those inputs in order for it to set appropriate IoT charging plans based on the resources consumed by each IoT device type and application.
How to keep an eye on IoT
The networking behavior of an IoT device type might also change especially after a software upgrade or the introduction of a new feature, causing a negative impact on the network resources consumed. The operator can work with the IoT device manufacturer to ensure that the device is friendlier to the network. The figure below shows an example of a signaling spike by an IoT device type after the introduction of a new service. The operator worked with the manufacturer to ensure that device behavior was back to normal without compromising the newly provided service.
In some cases, optimization might be needed in real-time; for example, to reschedule the upgrade of IoT devices connected to congested cells, or to throttle the users with anomalous behavior impacting the QoE of other users connected to the same cell.
Nokia Cognitive Analytics for Mobile Networks provides deep insight into the network to help in planning and optimization of the network to facilitate IoT — as well as enable operators to take actions immediately for the benefit of all network users.
Check out our Cognitive Analytics for Mobile Networks infographic for a handy summary on how Nokia can help with a wide variety of network challenges.
Share your thoughts on this topic by replying below – or join the Twitter discussion with @nokianetworks using #Analytics #BigDataAnalytics #IoT
1 As reported by Strategy Analytics in its 2017 report on IoT Cellular Device Annual Sales by Industry Vertical