Smart Cities: Challenges for Network Infrastructure

The concept of the smart city is growing in popularity in forward-thinking cities across the world. With automated systems collecting and transmitting data from more data sources than ever, everything from road intersections and surveillance cameras to park benches and rubbish bins can be monitored to identify problems, reduce unnecessary maintenance and make the most efficient use of city resources.

To support this wide array of devices and systems that will all need to be connected to a command and control centre where data can be analysed and decisions made, network connectivity that is fit for purpose is essential. The key part about this is fit for purpose – with such a wide array of devices, no one network technology is an ideal fit for all of them. Let’s look at some examples:

1. Video surveillance. This application generates a high amount of data, even with modern video compression technologies, thanks to the high image resolution and frame rates of modern cameras -coupled with their need to be recording at all times to properly serve their purpose. This rules-out the use of Low Power Wide Area Network (LPWAN) technologies, and with the exception of very remote or mobile deployments, also excludes cellular technologies to allow for the maximum number of users to utilise their expensive licensed spectrum.

These factors mean other technologies are needed to effectively support video surveillance on a city-wide scale. Beyond the need for a consistent high data rate, another consideration is the level of deployment of video surveillance desired across a smart city and the cost of that deployment. Laying and installing cabled network infrastructure is expensive, and doesn’t offer the same kind of deployment flexibility and speed that wireless networks do.

The ideal technology to support video surveillance in a city-wide setting is fixed wireless access (FWA), as it combines the ease of deployment of wireless technologies with the necessary uncontested bandwidth to support its continuous traffic demands. For some video surveillance deployments, such as those on moving city buses or trains, a combination of cellular connectivity and local storage with the contents being uploaded to central storage when the vehicle is at a station looks promising.

2. Sensors. Unlike video surveillance, sensor devices produce a small amount of traffic, often only heartbeat signals and the occasional status message. However, like video surveillance devices, sensors benefit greatly – even more so – from the flexibility in deployment that wireless networks provide. With many sensors being battery powered, low-power wireless connectivity is essential.

One consideration is how many concurrently connected devices a network can support. This is really a different problem to the amount of bandwidth the network can provide at any one point in time; instead, it stresses the part of the system responsible for identifying the state and location of connected devices, such as the location register in cellular systems.

To support these requirements of sensors, dedicated LPWAN technologies such as LoRaWAN are an option, as are more traditional cellular networks such as LTE. A big part of this decision comes down to who the network is being deployed and operated by – the city itself, or a traditional communications network provider such as AT&T for example.

In the former case, LoRaWAN and other sensor-targeted technologies are likely to find success as they provide lower deployment costs due to not having to use expensive LTE infrastructure. If an existing cellular carrier is chosen to provide network connectivity however, it typically requires only an upgrade to the existing cellular infrastructure covering the area to support these new devices.

3. Homes and businesses. When we talk about smart cities, it’s easy to get caught up in the newer and more exciting aspects such as sensors telling the sanitation department exactly when a given rubbish bin is full and ignore the fact that many people, even in developed cities today, do not have satisfactory internet access at their home or business. Sure, this number of people has been falling thanks to the increasing penetration of broadband internet technologies, but it still exists.

For long-life static locations of high value such as houses and offices, the most sought-after connectivity is from fibre. With its very high achievable speeds and lack of local contention, fibre is the technology of choice in many urban areas to connect buildings to the internet. However, the cost of deploying it can be prohibitive. Where deployment of fibre is challenging, FWA is also a useful alternative given a deployment which avoids or filters the interference prevalent in common FWA frequencies in dense urban areas.

Between these three, we can see that the potential number of networks deployed across a smart city is larger than might be expected. As they serve fundamentally different applications, it doesn’t look like a single network today is capable of providing satisfactory service for all of them. The closest, thanks to its combination of theoretical high-bandwidth and support for mobile wireless devices, Is LTE or a further evolution of the cellular network.

But, the limited spectrum these networks operate on and their existing use for connecting smartphones and tablets has them stretched to the limit in many dense urban areas today, without the additional load of supporting smart city infrastructure.

So, what does this mean for those thinking of how to architect and deploy networks for the smart city? It means that a heterogeneous network architecture is the order of the day, with all of the additional flexibility and operational complexity that this brings. Interestingly, it may be that the data received from these disparate networks needs to be combined in the centre of the network to be analysed on a regular basis – for instance, data from a video surveillance system being cross-referenced with a damage sensor on a traffic light to determine if the damage occurred by accident or as the result of a malicious action.

Ultimately, the network used must be decided by the applications it is delivering. Without a one-size-fits-all solution, the choice of network technology has a direct impact on the effectiveness of the network and on the smart city as a whole.