Why is IoT locked in 'proof-of-concept hell'? Stakeholders don't talk to each other, and return on investment is hazy Some Internet of Things projects are stalling, trapped in what analyst Canalys terms as a "proof-of-concept hell," though it believes the limiting factors are cultural rather than technological. On the private equity front, funding for IoT has crashed with just $500m spent on investments in the first three quarters of 2020 …
Wait, are you saying you would have to be an id-IoT to have an internet connected toilet (1), cock lock (2), "smart" lock (3), nanny cam (4), or similar device? I don't understand, what could possibly go wrong?
(1) https://www.digitaltrends.com/home/best-smart-toilets-ces-2019/
(2) https://lockthecock.com/products/cellmate
(3) https://www.cnet.com/news/have-a-smart-lock-yeah-it-can-probably-be-hacked/
(4) https://www.npr.org/sections/thetwo-way/2018/06/05/617196788/s-c-mom-says-baby-monitor-was-hacked-experts-say-many-devices-are-vulnerable
The article isn't talking about IoT in the sense of connected baby monitors and the ilk (rushed out by companies with poor security and unnecessary data grabs, unnecessarily dependant upon an unreliable web service), but of deployment of devices inside companies - asset tracking, door access etc. Without a link in the article to the rest of what the analyst was saying, it's hard to know exactly how they were defining it, but they did appear to be focused in business applications. (The original definition of IoT wasn't that things were connected, but that each 'thing' had a unique identifier. )
To say it isnt proven is not true - asset tracking (barcodes on equipment, for example, fulfil the original definition of IoT) and remote sensing (which would fulfil more contemporary definitions of IoT) is comon in industrial settings.
That's the problem with IoT, in the Enterprise.
Companies can have a lot of devices in their buildings. Fire alarms, phone systems, door entry systems, inventory management systems, cameras. Previously these may well have used their own physical networks. Security on these network s was nonexistent, but that didn't matter as you likely needed physical access to the network to get anywhere near the network. entirely separated from the outside world. Now, these systems are increasingly using existing networks, primarily Ethernet and Wifi, with sometimes quite powerful computers running them.
OK, so the Processor running your door entry readers isn't going to be able to run the latest AAA game, but it's likely got enough power that it can run a lightweight network sniffer, and if not properly put behind a firewall, likely able to send the data it has sniffed elsewhere. Because people see a door entry reader, fire alarm or camera, they don't see it's got a relatively powerful computer built in, so they forget them.
OK, systems like that can be hidden behind a firewall, and protected to some extend by NAT.
The problem comes because some of those systems need outside maintenance. Various companies have chosen to save money by supporting these systems remotely, so they need to be accessible to the outside world. Even if a VPN is needed to do it. All the IoT equipment and networking equipments/software used to connect it can have vulnerabilities, which can be exploited.
If the benefit of some huge investment in IoT 'stuff' is so unclear, or so diffuse throughout a business that no-one can become passionate about it, then no-one is going to champion it.
Proof of concept usually limits itself to "is this technically possible?". The problem is "Is this adding sufficient value?" Until one (or possibly two) stakeholders can say "yes", then we have a Mexican standoff, with everyone waiting for someone else to pull the trigger.
Alternatively, if your IoT thing only becomes valuable once everyone has fully invested in it, then - short of the CEO becoming an evangelist overnight - no company will risk adopting it.
Immediate and localised reward is necessary. "What do I get for investing in this gadget for my department?"
Everything that I've read or experienced about IoT applications has them being used primarily as a vehicle to enable pay-per-use type models, if not directly then indirectly through upgrades and enforced obsolescence due to dropped support. The industrial world has implemented IoT type applications for decades using efficient, standardized protocols with devices having well documented features. Common sense would suggest that IoT would build on this body of experience but instead its reinvented the wheel, building its entire model around complex protocols, inefficient data transfer mechanisms and proprietary applications. This model seems to be an extension to the smartphone ecosystems and wihile it might work for phone users (debatable) it doesn't work at all well for sensor and control applications. (Its the same sort of business logic that builds an "IFTTT" -- "If this, then that" -- application environment that completely ignores PLC/Open.)
For IoT things to take off outside of niche applications they've got to be simple, striaghtforward, reliable and open. They also need to be secure but not in a general purpose computing sense -- they do not need to communicate with third parties although depending on the design they may respond to a limited set of third party queries. Above all, they need to be able to be integrated into a system along with other vendors' equipment.
The phrase "Internet of Things" is a deliberate misnomer, a marketing ploy by data-harvesting companies to convince people to connect their networks of Things to the Internet so that those companies can harvest yet more data which you (or your Things) create.
Consequently, vendors are focused so single-mindedly upon convincing people that said vendors should have access to your Things and their data that they largely ignore the much more important issues of security and utility. Honestly, who needs that crap?
When they shift their focus creating a local network of Things which is useful and secure, and perhaps providing the user access to those Things via the Internet, maybe the public will show some interest.
"Many IoT projects have "failed to get beyond the proof-of-concept stage," Ball said, but the reasons for this are due to "internal politics"."
Not really. Many of them get to the proof of concept stage and the run head first into the reality that they're not actually useful. There absolutely are some interesting uses for IoT-style systems, and those are the ones that actually get used. But there are far too many things along the lines of internet connected fridges and internet connected chocolate teapots, and no-one actually wants any of those. It's nothing to do with internal politics or any nonsense like that, it's just people coming up with ideas that they think sound cool, building a proof of concept, and then getting a bunch of blank stares when they actually show it to someone. In fact, he goes on to say almost exactly that, if you read between the lines:
"We see successful IoT projects focus on clearly showing improvements in productivity, customer experience, compliance, and safety, as well as enabling new revenue streams."
Successful IoT projects focus on actually doing something useful, unsuccessful ones, by implication, don't. That's not a cultural or communication problem, it's the system for once actually working as intended.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
