How APAC organisations can harness the power of IoT

Use cases and considerations

IoT is “a lot broader than most people think, and a lot more accessible,” observed Bobis.

Examples include connected cars, fridges, utility meters and even smartphones. The data collected from such devices is being used in various ways, such as providing better routing of trucks in the face of road closures following major floods, the delivery of real-time updates, and warnings of possible water leaks.

And in Australia, toll road operator Transurban has been able to use data from vehicles about swerving and braking to halve the number of crashes on one of Sydney’s major roads, she said.

Perhaps more prosaically, Micron uses IoT technologies to monitor the operation of equipment within its factories. Fong said vibration sensors can help determine whether fans are operating correctly or if they will soon need to be replaced, and cameras monitor robotic equipment to ensure that parts are being placed correctly and the tools are in good order.

A sensor may be able to take measurements every second, but the communication and storage costs may outweigh the value of such finely grained data

In considering use cases for IoT, Kuppusamy pointed out the importance of understanding the business value – or indeed any other type of value – of a proposed IoT project. “It’s not about monitoring something or sensing something, it’s contributing to a tangible outcome” such as bettering a community, human lives or the environment, or helping achieve a business goal,” he said.

Organisations considering IoT implementations should also be aware of two particular barriers to adoption that are human rather than technological, Bobis noted.

First, there is still a need for more education about IoT, as “it can be a bit nebulous and a bit pretentious”, and that can put people off finding out more about IoT in general or specific IoT solutions.

Second, “innovation does not equal acceptance”. If something has been done in a particular way for decades, “you can have this amazing solution which does provide value”, but it won’t necessarily be accepted, she said.

“My favourite example of this is seat belts. When they first came out, people were so against the idea,” she said, adding that this was to the extent of arguing that you were more likely to die in a crash if you were wearing a seat belt – it took 25-30 years before they became mandatory.

So, organisations will need to adopt an innovation culture that allows risk taking, as people need to be confident that they won’t lose their job if they try something and it fails, Bobis said. Kuppusamy agreed but pointed out the importance of managing failures when they do occur.

Challenges with IoT projects

Given that IoT is still a relatively new field, Kuppusamy’s warning that “there will be some challenges” shouldn’t come as a surprise. Those challenges can be encountered in various aspects of a project.

An obvious challenge is data volume. A sensor may be able to take measurements every second, but the communication and storage costs may outweigh the value of such finely grained data. And so, it might be more sensible to aggregate the data in some way, suggested Kuppusamy, such as averaging it over longer periods, or to simply take readings less frequently. However, it is sometimes important to capture data in real-time to detect crucial events when they occur.

There are also challenges associated with the environment in which sensors are deployed. For example, Kuppusamy said soil types could be an issue for a project aimed at collecting rainfall and river flow data at mining sites. This is because the ground is usually soft in salt mining areas, but very hard at iron ore sites, so securing the required sensors is non-trivial.

In remote areas, IoT implementations typically call for satellite communications as well as hardened devices that can function reliably over long periods without human attention. That reliability may also be required in urban areas for economic reasons, such as the cost of replacing batteries.

As such, Kuppusamy suggested taking a value perspective. A $10,000 device is unlikely to be justifiable for most projects, no matter how rugged or durable it is. On the other hand, having to manage devices, whether it is repairing them or changing batteries, is also expensive. In most cases, better results come from installing more sensors rather than better sensors.

Privacy and security are also important issues that do not always get the attention they deserve. Organisations are becoming more security conscious, said Tan, and it is becoming normal to assume that something may go wrong, so steps should be taken to prevent and mitigate security breaches.

One of the coolest things that's going to happen in IoT is for transport planning applications around decarbonisation

“Talking about it is really important,” he observed, noting that part of the problem is that manufacturers may not be inclined to spend the money needed to secure cheap devices – although Singapore is incentivising IoT manufacturers to adopt a security-by-design approach.

With application programming interfaces (APIs) commonly used by IoT applications, the fewer APIs used and the simpler they are, the easier it is to secure them, suggested Kuppusamy. He recommended keeping things simple, and then taking steps to protect the simplest and most vulnerable APIs.

Another issue arises when actuators – rather than just sensors – are added to the mix. “If you go to a restaurant and there are robots serving food, the question is, if something goes wrong and the robot pours hot soup on somebody, who takes the liability?” asked Kuppusamy. And if your home security is based on IoT, “the last thing you want is to be locked out of your own home”.

More generally, there is the question of whether an IoT system should be able to respond autonomously, or if there should be a human in the loop. “If your confidence levels are 90% and above, the capability of your solution is pretty good. But if your confidence level is at 50%, then obviously the human aspects to provide the contextual information is quite key,” he said.

Industry landscape

Tan noted that the IoT landscape is highly fragmented, with suppliers playing in different parts of the IoT technology stack.

Kuppusamy pointed out that IoT suppliers may need to work on projects alongside their competitors if they do not have the necessary expertise in all aspects of the technology including devices, networks and applications. But as Tan pointed out, there is a tendency for IoT suppliers to expand their technology stacks over time.

IoT should also be regarded as an ecosystem, according to Fong, one that requires expertise in hardware, software, cloud systems, analytics and mathematical models and security. “It’s a team effort,” agreed Tan.

That said, Kuppusamy stressed the importance for IoT technology providers to specialise in an area or industry: “There is a temptation to spread your net widely as the same sensor can be used for different applications in different industries, but you need to be able to service your customers effectively. Each project needs to help move the company in the direction it has set.”

Future developments

Not surprisingly given the companies they work for, the panellists see different opportunities in the IoT market in the next five to 10 years.

“I think one of the coolest things that’s going to happen in IoT is for transport planning applications around decarbonisation,” said Bobis, helping to answer questions including: ‘How can we get people to get an electric vehicle instead?’, ‘How can we put in enough infrastructure so that people who buy these vehicles can charge them on the street?’, ‘If everyone’s trying to charge their car at the same time, what is that going to do to the electricity grid?’ and ‘What about the bidirectional transfer of energy between the car and the grid?’.

Kuppusamy, too, sees opportunities in the utilities sector, and Space Angel is working with both the electricity and water industries.

Fong’s team is doing a lot of work in the sustainability area, collecting data about the use of water and other consumables, and then developing models to reduce that consumption in manufacturing plants. Among the questions being addressed are why different tools carrying out the same processes are consuming different amounts of water, he said.