Your browser is not supported. Please update it.

4 August 2017

Canonical survey finds IoT skill-gap, internal disputes, main IoT challenges

Linux-titan Canonical has published the results of an industry survey that it commissioned, finding that many businesses are still not on board with an IoT project – concerned with ROI and proven business benefits. The main focuses of the survey were identifying how the IoT can be monetized, what skills are required, and how to address fundamental security concerns.

Some 361 ‘IoT professionals’ responded with their experiences and opinions, with 79% noting that the security threat for the IoT was not being given enough weight by the press. Recruiters are noting something of a skills gap, but IoT projects are still trying to prove their worth ahead of deployments – with quantifying the business benefits still the main concern.

Citing a McKinsey figure of a $900bn current market value for the IoT, the report notes that many businesses are still grappling with how to integrate the new technologies into their operations. It is well worth your time to peruse, but we’ll dive into the main points below – monetization, employee skills, and security.

For expected benefits from the IoT, 26% of respondents primarily pointed to a better quality of life, with 25% noting greater business insights and another 25% pointing to new service and categories. Some 18% said they expected reduced opex, and just 4% believed it would reduce capex.

As for the most immediate challenge faced by the IoT , 53% pointed to a lack of a clear use and the problem of quantifying a return on investment (ROI). Some 45% pointed to device security and privacy, with 40% citing a lack of IoT infrastructure. Lack of budget claimed 34%, ecosystem integration scored 29%, device management and long-term support was on 26%, with internal organizational resistance at 25%.

In last place was ensuring regular updates, on just 12%, which suggests that installers are pretty confident that they can keep things patched – once they can justify the existence of the project in the first place, and secure the funding and successfully integrate it into the existing systems.

When asked about what was needed in order to encourage enterprise IoT adoption, the survey found that 34% said quantifying business benefits, 24% cited an improved understanding of the benefits of the IoT, and 17% said improved security. Just 11% said better integration with everyday systems, 7% pointed to better data use understanding, only 5% think device costs have to be lower, and ‘localized infrastructure’ was a concern to only 2% of respondents.

Those results are pretty encouraging, with the smaller concerns suggesting that the IoT technologies are already in place – enough that they can be turned into viable business projects. But the concerns that scored higher suggest that the bigger issue is in convincing decision makers in the business to get on board and support a project – although ROI concerns are very valid, and blind faith is rather risky.

The survey found that 55% believe that hardware sales drive their profits, and therefore that the volume of hardware the IoT will require will mean increased sales. However, the survey correctly notes that there will be considerable commoditization pressure in the IoT, which will eventually crush the margins out of established hardware markets. Early-stage devices will enjoy nice margins, but soon they will be completely commoditized, if they ever reach scale.

As Canonical puts it, there are two choices for hardware makers – bespoke premium hardware targeted at less price-sensitive market niches, or an approach to fitting the software or application stack on general purpose single-board computers (SBCs) or systems-on-chip (SoCs).

When it comes to monetizing the IoT, there appears to be more industry consensus. Some 78% said they expected to earn their keep through value-added services and maintenance, with 57% looking at hardware rental models, but 55% envisioning a one-off hardware fee. Another 55% said on-going software and security fees would be the key, with 54% citing consultancy and deployment fees. Somewhat adrift in last place was the consumption of IoT services, at 40%.

These sentiments are in line with what Riot has seen in the wider industry. The real value in the IoT is at the data layer, and the edge-devices are simply a conduit to bring that data to a central application – which can then turn these data streams into something useful that can turn a profit. Because the key is data, AI-based analytics technologies are going to be of great importance to these enterprise players, as a means of efficiently sorting and acting on the vast quantity of incoming data.

Turning now to the issue of a lack of business knowledge and skills, and the need to hire the right kinds of skills to adopt an IoT project, Canonical notes that given the IoT’s relatively new status, “many business leaders have found themselves running headfirst into a set of technology and business challenges that they do not yet fully understand. What they need is a new generation of talent with the knowledge and skills to navigate the current Wild West that is the Internet of Things.”

When asked, some 68% of respondents said they were struggling to hire employees with IoT skills. In terms of the areas that are most difficult to hire for, data analytics and big data was in first-place on 35%, with 33% citing embedded software development, and 32% on embedded electronics. Close behind was IT security, on 31%, followed by AI on 30%, and then trailing on 17% are cloud software development, with automation on 14%, and robotics on 13%.

It isn’t clear if AI’s 30% is due to the lack of demand for positions, or a lack of staff, as it is such a nascent area. We’re sure that a new crop of graduates will soon begin swelling the recruitment pools, but an increase in demand for these types of skillsets could see recruiters struggle even more to fill those positions.

AI could take a path similar to what is apparent in the demand for cloud development skills, where only 17% of recruiters are struggling, but it is possible that it could surge much higher – if AI-based data analytics tools become vital to improving profits. It may very well be a good time to be an AI developer.

IoT Security was the third and final major focus of the survey, with Canonical pointing to Mirai and WannaCry as examples of the kind of concerns facing the industry. Of course, Ubuntu would like to make the case for its Snap containerized applications as a bastion of security, for enterprise devices looking for secure field support and upgrade cycles, but hard-coded passwords are still a thing that needs to be countered by manufacturers and OEMs.

The survey found that 21% of respondents think that the security threats have been ‘hyped up’ by the media, however that other 79% think the press is presenting an accurate or underplayed picture of the IoT’s security issues. “The genie is out of the bottle as far as IoT security is concerned. As hackers get ever more interested by the swathe of poorly protected IoT devices, it seems likely we’ll see more such attacks take place, with malicious agents utilizing swarms of IoT devices to compromise commercial entities.”

As for corporate responsibility for security, 57% of respondents thought that there would be joint responsibility between the businesses involved in deploying an IoT system – to keep everything up to date and secure. Encouragingly, only 3% said they didn’t know who was responsible for security, while 11% pointed to the IT department. Some 12% said the manufacturer was responsible, and 14% answered that it was the remit of the vendor or supplier.