Brian Back, founder and CEO of Radio Data Networks, discusses his concerns about the use of IoT on third-party carriers for mission-critical applications, and TETRA as a backbone for remote monitoring solution the Critical Network of Things

Not a day goes by without mention of the Internet of Things (IoT), with the usual players seeking global domination in a Klondike-like gold rush to capture their share of a market that analysts estimate already to be worth many hundreds of billions of dollars and will be, according to some, worth as much as $7 trillion by 2020.

Who could doubt these figures if looking at the total potential connected market as a whole – that is, the IoT package with its host from autonomous vehicles to intelligent personal assistants. However, if you consider the IoT part on its own – say, just the sensor, the data delivery and hosting – then these estimates must be questioned.

For at one end of the scale, hardware providers are talking of devices at just $20. In the middle, network operators are already offering connectivity for just $5 per annum. However, the ultimate client, the public, are already of the belief that data should be free. So, before the IoT market is even in full swing, the commercial squeeze is on from both ends.

Perhaps this explains the ‘infant mortality’ rate – according to a study from Cisco, 60 per cent of IoT initiatives stall at the proof-of-concept stage, and only 26 per cent of companies have had an IoT initiative that they considered a complete success. Even when projects were completed, a third weren’t considered a success.

Coming from a remote monitoring, M2M and control background, spanning over three decades in electricity, rail, water and the environmental sectors, what sends shivers down my spine is the level of naivety among the latest generation of potential customers and perhaps, worse still, the openly irresponsible promotion of IoT for these and other potentially mission-critical services. Browse the internet and the evidence is right before you, with images of trains, aircraft and electricity distribution networks all being promoted as candidates for monitoring (and control) using IoT technology.

To list all my concerns over such a broad-brush, one-size-fits-all IoT proposition, it would take up more than the space allocated for this article; however, if pressed to choose, my top three would be ‘responsibility’, ‘testability’ and the dependence on the internet.

Taking responsibility first – if something goes horribly wrong, who will take the rap? With many IoT solutions, the chains are often very convoluted and as a customer you could soon end up chasing your own tail not knowing what has gone wrong: the sensor, the IoT device, the network, the sim card, deliberate jamming, interference, network congestion, a failed network hub, service outages, a denial-of-service (DoS) or distributed denial-of-service (DDoS) attack, the fog, the cloud or the data application. Certainly, based on the aforementioned economic models, it would be highly unlikely that any IoT provider would enter into a service-level agreement or could afford to fund an adequately staffed 24/7 call centre to deal with any if not all of these issues.

Accepting the inevitable that service standards will be low, then my second point – testability – becomes more important than ever. As many new to IoT/M2M will soon find out to their frustration, and even the potential early demise of their business, you can easily waste time and money driving all over the place to access suspected failed devices only to find out the issue lies elsewhere. However, to get to that stage, the first challenge is determining whether or not your device is working. Take cellular IoT; the big issue here is that unless the messages go all the way through the network to the cloud, there is little way that you can conduct an on-the-spot diagnosis. Listening on a radio scanner for messages will get you nowhere, and the networks are such that to listen to a message ‘sniff’, as it is called in the trade, and know that it is yours is virtually impossible, and that applies to the vast majority of SRD (short range device)-based IoT protocols too, from Wi-Fi to LoRa. Further, having detected a signal at the device doesn’t necessarily mean it is reaching the base station. Amusingly, I have heard of sorties to reach devices located beneath roads where, surprise, surprise, during test and with the manhole cover lifted slightly to allow the test cable to enter, they work – only to find that they fail miserably the moment the test cable is removed, the cover is closed and it rains.

Moving onto the internet – this is an ongoing issue that can only grow. Already there are numerous stories concerning IoT devices being hijacked by hackers with malware, trojans, bots and the like being installed, even ransomware cases. Ransomware concerning IoT could also take two forms: first, when the devices themselves are hijacked; or alternatively, when other devices are used to launch DoS or DDoS attacks, because denial of service can be just as damaging to a mission-critical IoT application as a network outage.

In summary, for IoT to have any chance of being mission-critical, we need to instil accountability, establish responsibility, take ownership and get the end-users to simply understand that they get what they pay for.

The birth of the Critical Network of Things
Fortunately, I am not the only Luddite, as I have often been called, who dares to ask the “what if” question. In 2015, following a chance meeting at Ofcom in London, I found myself within months sharing thoughts across a table in Bruges, Belgium with an equally likeminded Philip Vercruysse, the founder and CEO of Entropia. Over the course of several cups of coffee and a visit to Entropia’s London Pub (yes, they have a quintessentially British pub tucked in the corner of their Bruges office with Chesterfield-like seats), we coined the phrase the Critical Network of Things (C-NoT). We selected Entropia’s international TETRA network because of its proven resilience as a backhaul. Simultaneously, we formed a joint-venture between Entropia and my UK-based business, Radio Data Networks (RDN), through which the latter’s RDNET1000 wireless sensor technology and data gateways would be converted to support the TETRA short data service (SDS) protocol. Not a bad day’s work and well worth the sacrifice of a few cups of coffee and a bottle of Belgian (and Dutch) beer!

Prior to this meeting, RDN had already established in the UK what is probably Europe’s – if not the world’s – largest sewer monitoring network using licensed PMR radio technology, delivering at that time more than 100 million messages in virtually real time to the water utility sector alone, but via physical gateways into the utilities’ existing infrastructure (outstations). The adaptation of this technology was the obvious route to proof of concept for C-NoT, with the gateways modified to packetise and push the data across the TETRA networks back to ComBus, Entropia’s mission-critical hosting platform installed within two military-grade servers (soon to be three).

Within a few weeks, we had a proof of concept running, focusing initially on wastewater sewers and overflows, then moving onto water supply, delivering hourly temperature and water-level measurements from remote strategic water infrastructure assets such as boreholes.

C-NoT is a hybrid system, which uses a narrowband wireless telemetry unit at the borehole, to interrogate the sensors using the SDI-12 protocol with messages relayed to the TETRA data gateway where it is both archived and forwarded over the TETRA networks using the data service. During the trial, RDN’s RDNET1000 protocol proved ideal as its lean, bit-orientated structure permitted not just single but multiple messages to be accommodated within the relatively limited 140-byte packet constraints of the standard TETRA SDS payload. (Try using a protocol such as JSON – JavaScript Object Notation – and you would be lucky to get a single message across).

Mission accomplished – a unique combination that could both deliver virtual instant messaging (tamper alarms – when you break in, it sends an alarm message) while being able to achieve the Holy Grail of a 10-year-plus battery life at the sensor. Not forgetting testability; the gateways also have the ability to be interrogated on demand, capture the received signal strength indicator (RSSI) of incoming messages, and conduct background scans for noise and channel interference.

The pilot batch of TETRA gateways

From an outsider’s perspective, the most attractive part of Entropia’s C-NoT proposition must be the ability to close the ‘responsibility gap’, as by using TETRA they have achieved true control, the desired levels of testability and visibility from the tip of the sensor to the screen.

One year on, we have added multiple new dimensions to the C-NoT, offering spatial and frequency diversity on the TETRA gateways and the facility for the direct attachment of local wired sensors. We have also introduced the ability to create dual and even triple redundant networks by supporting DMR, PMR and satellite in parallel to the TETRA backhaul.

Finally, with last week’s acquisition by Entropia of the Mobitex networks in Belgium and the Netherlands from Ram Mobile Data, it will be just a matter of days before the Mobitex networks are also carrying mission-critical C-NoT data.

Author: Brian Back