The IoT offers such a wealth of opportunity that you need the courage to work with others in order to move faster, argues one of the industry’s leading lights, Drew Johnson, vice president of engineering at California-based service provider Aeris. Report by Nick Booth, freelance IT & communications writer.
IoT Now: What trends have you seen in your time in the industry?
Drew Johnson: There have been so many in the five years I’ve been at Aeris. Just in that time, of course, we’ve stopped calling it the Machine-to-Machine industry and now everyone says it’s the Internet of Things and all the big companies have gotten into this space.
According to our recent study of enterprises which we released in May 2016, the three main motivations cited for IoT are to boost revenue (noted by 14% surveyed), create competitive differentiation (17%) and monitor products (23%). The number of companies that have a specialised IoT department is still a minority (40%), but that proportion is growing.
The newest emerging trend, which we started talking about at the end of last year, is for companies to take an ‘IoT-First’ strategy, where their product is now designed, from the outset, with IoT as a primary consideration. Just as ‘Mobile-First’ changed the mindset of web designers, so they no longer imagined the systems they were creating being used in a fixed internet setting, so has IoT forced a redesign of all aspects of the product.
This is exemplified by the cases of Nest and Honeywell. Whereas Nest’s ‘IoT-First’ approach re-imagined the user experience and interaction with an ecosystem, Honeywell just bolted on its IoT capability as an afterthought. Nest created a system that can make pro-active, intelligent interventions on your behalf if you are not at home. It’s simple but powerful. Nest also connects with an ecosystem. It can interact with the power companies and even with other devices in the home.
IoT-First means making IoT a design priority, so you start out with a fresh set of plans. We see this also in the connected car space. Granted, cars have been connected for 15 years, but the original concepts were limited.
These days, designers are re-imagining events at every stage of a car’s life, from manufacturing, through shipping, the time at the car dealership, then ownership and warranty issues, all the way through to scrapping. Each one of these processes can be made more efficient – and precious resources can be saved – by designing them with IoT in mind. Just as Nest has done this well on the consumer side, Tesla has set the example in IoT design for the connected car.
Now, other companies are starting to see this approach as the standard to which they should aspire.
IoT Now: Can you clarify the confusion created by the wealth of IoT platforms?
DJ: You are right. There are so many IoT platforms out there that it can be very confusing.
We tend to look at the anatomy of any IoT system as having seven core elements that need to be harmonised. These seven elements consist of:
- The “thing”, the device itself
- The connectivity
- The data being exchanged
- The analytics of that data
- The visualisation of this information
- The creation of value, meaning the monetisation
- The security
Though we like to highlight security as an independent element, it should be endemic in all the other elements of the IoT too. Yes, overzealous application of security disciplines can slow projects down, but we should bear in mind that a system is only ever as secure as its weakest link. So anyone could take one component of a system and pick the rest apart.
For example, in fleet management it has been proved that in some cases a simple SMS message has the potential to do a lot of damage. Through a text message you can pinpoint a machine’s location, and it offers the potential to redirect information. So that’s a big risk, created by a small element within the grand scheme of things. That’s why security (across platforms) is hard for many companies, as they can build a system but not a holistically safe one.
There are platforms available for every element in IoT. Meanwhile, the public clouds are also becoming increasingly influential in the IoT. There is still a lot of complexity in leveraging the public clouds today, but they will definitely be the right answer for some companies and will play an important role over time.
To lessen this confusion, I would offer this advice. Try to select a platform that cost-effectively delivers most of what you want and is in harmony with the public clouds. Concentrate your resources on the elements that will provide true value and differentiation to your customers. Those elements are typically the customer experience. You will save a lot of time and money by allowing others to deliver what is in the middle.
IoT Now: What type of analytics are most important to start with?
DJ: The best way to tackle the challenge of analytics is to start by thinking about the answers you need and how quickly you will need them. There are three types of analytics challenges, based on whether you need to look at real-time intelligence, batches of information or longer term trends from deeper analytics.
For example, the safety and security analysis we conduct on airplanes falls into all three of these categories at different levels. It’s not just about, say, potential crash data, but long-term analysis of trend-related intelligence, which can give historical information on components and contextualise that with recent reports on the current state of play. This data allows us to identify a component that is worn and needs replacing, so we can prevent a problem from becoming bigger, replace the component and get the new equipment tested and in place for the next flight.
For an effective IoT analytics platform, there are three important interfaces to consider. There needs to be an interface for the data scientists to create their long-term prognosis. There also needs to be an interface for the data engineers to get high-performance access for the online use cases. Finally, there needs to be a way to let the business managers use the intelligence and visualise ways to use it to run the enterprise more efficiently.
In May 2016, we completed a survey of enterprise IoT in global corporations. We found that the biggest priority among the survey group was analytics, with 21% of those questioned saying IoT data analytics capabilities are important to their IoT service. More people nominated this category than any other category followed by device management.
IoT Now: What advice would you offer from a technologist’s perspective?
DJ: Many predict that IoT will soon impact every single enterprise. My advice depends a bit on whether the enterprise is acting as a consumer of IoT or creates products and will be a producer of IoT.
When the enterprise is a consumer of IoT services, the key to success is choosing services that are open and can be integrated into other services. For example, an enterprise selling apples may have farming equipment and trucks and cameras at different points of the production process, but unless they can tie them all together, they won’t be able to get consistent (and most cost-effective) management as those apples proceed from the orchard to the warehouse to the shops.
Our survey of all the professionals who are involved in deploying IoT programs, across a range of industries, identified that integrating software and platforms is most popularly nominated as the most complex task. One fifth of those surveyed mentioned this first. Only security (which was nominated in 18% of cases) came close, but security and integration are co-dependent.
When acting as a producer of IoT, enterprises must be able to tie their services into those created by others. So they must they must choose technologies and platforms that readily tie into the ecosystem. We are actually focused primarily on the IoT producers. We want to help them take the journey from unconnected product to connected product with an IoT-First approach.
IoT Now: Which innovation excites you most about the future of the IoT?
DJ: As a technologist, I’m super-excited about where we are right now. Overall, it’s still early in the IoT evolution. I think there are several drivers that need to continue to emerge in order to take IoT to the next level. We have to have low-cost devices and sensors. That area has benefited already quite a bit from the sensors, modules, and processors being produced in mass quantity at low cost for mobile phones. We have to have low-cost connectivity. That’s being delivered by cellular in many cases but also by WiFi and new Low-Power IoT networks. We also have to have low-cost ways to analyse all the data. It’s been amazing to see all the open source Big Data technologies emerge over the last 6 or 7 years. Finally, we also need new ways to interact with products. Now, nearly everyone is walking around with a smart phone and there is a new emergence of voice interactive agents like Amazon’s Alexa.
I have to say that I really believe this IoT-First concept will create a lot of value for consumers, for product-makers, and even help the world save resources. We already talked about the Nest thermostat. It saves the consumer time and money and also helps the power companies be more efficient. Another simple example is a watering system that does not water lawns or crops when it rains or is about to rain. Imagine how much water could be saved. There are hundreds of thousands of such products that can be empowered to talk to each other and use a machine form of common sense.
IoT Now: What advice would you give to a company just getting started with planning their IoT programme?
DJ: There is endless scope for disruption in this industry. If I were starting out I would be very focused on how to re-imagine the product interactions around IoT and connectivity. Try starting from a clean slate without pre-conceptions on how users will interact. Focus also on how not just to use the connectivity to pull information out but also how to tie into other information as part of a bigger ecosystem. Once you have done that then it’s important to get something working quickly in order to get feedback and improve rapidly.
IoT Now: From your perspective, why are some IoT programs more successful than others?
DJ: We see those that struggle are the ones that try to execute in areas where they don’t have the expertise. If you don’t leverage the talents of key partners, you will likely struggle too. If you are not strong in all of the seven elements of the IoT, then it won’t work.
Make sure you do your own part very successfully and get the best you can from partners in the other areas. Nobody can be best at everything. We see a lot of enterprises fail at data management and analytics with respect to scale. We have seen companies struggle when they attempt to go it alone, and we like to help them to focus on areas where they have the maximum value.