The Internet of Things is starting to become more mainstream. Consumers around the world are buying smartwatches to count their steps, smart thermostats to control the temperature in their house based on whether or not they are home, and even refrigerators that can help them make their shopping list. Businesses are also taking advantage of IoT to increase productivity and efficiency. But in order for users to really gain value from these devices, the devices have to be able to send and receive data somehow.

“The ‘T’ in IoT refers to a wide variety of ‘things’ that usually only make sense to put on the ‘I’ (Internet) when that allows them to be integrated with other systems. In other words, IoT is as much about the Integration of Things as it is about the Internet of Things,” said Uri Sarid, CTO of MuleSoft, an integration platform provider.

The problem, however, is that integration involves the IoT device, data integration, back-end system integration, and third-party middleware integration, according to him. In a recent Gartner report, the technology research company revealed integration was a top challenge for businesses.

(Related: What’s in store for the next generation of the Internet of Things)

“Anyone can produce an instrumented device that has a cool widget, does some analytics, and puts up a screen for someone to look at and monitor. But the real art is integrating that,” said Benoit Lheureux, a research vice president for Gartner’s IoT research group.

Without the integration part of IoT, devices become useless, he explained.

Greg Gorman, director of developer ecosystem and cognitive analytics at IBM, believes the importance of IoT will only continue to grow and impact society because it provides convenience; makes a person’s life easier; the data can be used to make better insights and decisions; and because at the end of the day it is just fun.

Solving the integration problem
IoT integration has become a problem simply because it is expensive, according to Gartner. The firm predicts that through 2018, half the costs of implementing IoT solutions will be attributed to integration.

“There are still a lot of companies that don’t get that, so they aren’t investing sufficiently in integration,” said Lheureux. He went on to explain that businesses commonly fall into the problem of treating integration as a second-tier focus rather than a primary focus. Instead, they will put all their energy into choosing the right sensors and partners, and ingesting the data and analytics, but won’t put enough effort or resources on the back-end system integration. This is because integration isn’t as necessarily appealing as the devices, he explained.

In addition, it is expensive and hard to keep the physical part of IoT up to date, according to MuleSoft’s Sarid. “All of these devices rely on an integration to provide access or information and can span a wide variety of technologies, locations, operators and sensitivity levels. Often the data they can provide is vast, but it may be hard to transmit given the physical limitations of the devices and their environments, and making sense of it can seem like looking for a needle in a haystack.”

Ways to overcome integration challenges
• Have the business rationalize their integration strategy, according to Lheureux. There are three things he believes businesses need to ask in the early stages of their IoT planning:

  1. What are their fundamental integration challenges?
  2. What complications might arise if the scope of integration expands?
  3. How does IoT tax existing integration skills?

Lheureux also says businesses need to take on a pervasive integration strategy where they stop treating integration as a discrete problem, and start testing it as an overarching skill.

• Proficiency at application programming interfaces (APIs) is also a game-changer, according to Sarid. “Access to IoT devices or their controllers should be done through strategically designed and productized APIs to provide a well-defined ‘surface area’ for the IoT devices,” he said.

“Here security can be managed, integration can take place in a controlled manner, upgrades and versioning can be rationalized, and visibility can be provided. An API-led approach can then be put in place to provide the integrations to other systems that enable agility, require less maintenance and reduce expenses.”

Sarid adds that an API-led approach can also allow businesses to implement new technologies faster and adapt quicker.

The problem with an API-led approach is that the design may be too complex for device manufacturers. Instead, they should follow best practices, such as “creating APIs in standard, predictable ways with design languages such as RESTful API Modeling Language (RAML) and JSON. RAML creates a common language for talking about the API itself, making it easy to manage the whole API life cycle from design to sharing,” according to Sarid.

• Recognizing IoT integration is not just about IoT, according to Gartner’s Lheureux. For example, when you look behind the scenes of a simple Febreze Connect product, it involves cloud service integration, a mobile app, integration with the back end, and the home air freshener.

• Embrace open standards, according to IBM’s Gorman. Using something that is just intended to be in your company’s cloud or stack is a narrow-minded way of thinking and is going to stifle innovation, he explained.

“You have to think of it in terms of, ‘We are building these devices, we are building some technology, and people are going to use it in ways that we had no idea and never intended it to be used for.’ You need to make it as open as possible and in that way you are going to succeed,” Gorman said.

• Recognize the need to develop a new app development model, according to Rajeev Kozhikkattuthodi, vice president of product management for TIBCO Software (a software infrastructure and business intelligence provider). He explains that we are in a software development world where millennials and post-millennials will soon become producers of technologies like the Internet of Things. These millennials desire real-time and continuous feedback; however most IoT software development and design paradigms haven’t evolved in decades, he explained. “We believe designing app development tools for millennial and post-millennial developers requires a much more agile, conversational and adaptive experience,” he said.

Phase one of the Internet of Things was enterprise adoption of IoT. Now that projects are coming out and becoming smarter, we are heading into a new phase where integration needs to happen. “The future of IoT integration is nearly here because people are realizing that IoT doesn’t work without integration,” said Sarid.

“If integration is not prioritized as more IoT devices make it to market, consumers may not adopt these devices because they would likely add far more complexity to their lives rather than their promised simplicity.”

Integration patterns
When developing a product, if you can use a full Internet and web stack, then you should, according to Dom Guinard, cofounder and CTO of EVRYTHNG (an IoT smart product platform provider). However, that is not always the reality, and with all the standards and protocols out there, things can become difficult for a developer.

According to Guinard, there are three common integration patterns developers can use to bridge protocols to their applications:

  1. Native integration that allows you to transform your device to talk natively to the web. This is well suited for powered devices.
  2. Gateway integration patterns that allow you to communicate with a gateway, and to translate the message back and forth.
  3. Cloud integrations that allow you to connect with a certain protocol to the cloud, and then the cloud provides a web and Internet view on top of the device.

A new way of dealing with integration
One of the biggest challenges facing IoT integration boils down to complexity—complexity in terms of scale, and in terms of the IoT ecosystem, according to TIBCO’s Kozhikkattuthodi. He explains that with today’s IoT platforms, there is a vast array of choices in terms of protocols, implementations, and vendors trying to solve an interesting (but small) piece of the broader challenge.

There is also a radical new paradigm emerging with IoT. There are billions of devices coming online, and all the integration challenges coupled with the need to develop a new rapid agile app development model are posing obstacles to IoT integration.

To address this, TIBCO has introduced Project Flogo: an open-source lightweight IoT integration solution. “What we saw was there was an order of magnitude shift in terms of the sheer scale of problems and the complexity offered by that, but also a magnitude shift in user expectations around the application development paradigm. That was what really drove and inspired Project Flogo,” said Kozhikkattuthodi.

Taking a fresh look at app development from a user-experience standpoint, Flogo is designed to develop a smarter user interface with intelligent design agents. The project’s design bot provides a software agent to interact with developers. Users can converse with these agents as they undertake application design.

Flogo is also designed to provide a new debugging aspect. Its step-back debugger allows developers to “step back in time,” meaning they can visually develop and simulate sensor events, as well as go back and make changes without restarting the complete process. Kozhikkattuthodi says this is necessary with IoT because it is difficult to simulate real-world IoT inputs, such as temperature, air pressure, and location.

On the technology front, TIBCO designed Flogo to shrink down to the lowest and smallest form factor possible. “This is logical because we need to run on extremely small devices that are extremely resource-constrained,” said Kozhikkattuthodi. “It is one thing to run on an enterprise server, and one thing to run on a massive cloud infrastructure versus running on a tiny device that is out there in the field. A lot of focus has been on bringing the footprint down.”

However, it is not sufficient just to bring the runtime footprint down, according to Kozhikkattuthodi. Flogo also undertakes an edge-native design standpoint, allowing developers to design for extreme resilience. This allows devices to engage in troubleshooting from remote locations, and to be resilient to arbitrary hardware failures and issues.

Another tool for connecting the IoT together is called Node-RED. Node-RED is a homegrown IBM project recently rolled out into open source through the JS Foundation. Node-RED is designed to give developers a visual tool for connecting IoT devices, APIs and online services. “It is a really easy-to-use, graphical way of connecting devices and managing their data,” said IBM’s Gorman.

Node-RED features a browser-based flow editor, JavaScript functions, and a built-in library with functions, templates and flows for reuse. “It is not associated with IBM in any way. It will run on all different cloud servers and connect elements together from just about anywhere,” said Gorman. “There are thousands of contributed nodes to it that make it easy for people to communicate with devices, connect them together, and do really advanced data science.”

Focusing on the things
When you think of the Internet of Things, most people think of the Internet and connecting things together. But there is also the intelligence of things—bringing that intelligence to the device.

“If you can make things intelligent and not just connect them, you can suddenly create exponential value and transform your whole system,” said Timothy Butler, CEO of Tego (an asset intelligence provider). “We walk around with literally gigabytes of information at our fingertips that enable us to make a lot better, smarter and more efficient decisions, and it is the same thing with things. If you think about things as an extension of who we are, then having intelligence on those things becomes an almost obvious interaction.”

For instance, the app Waze is more than just a traffic monitoring system. According to Butler, it collects data at the edge. Without feedback at the source, such as data collected by driving down the road, the information wouldn’t be possible.

“As people are connecting things, what they are realizing is the different types of data they can have at the edge,” said Butler. “Edge computing is pushing the frontier of computing applications, data, and services away from centralized nodes to the logical extremes of a network. It enables analytics and knowledge generation to occur at the source of the data.”

Depending on the asset, asset intelligence management can help a user easily track, configure and update all the information coming from a device without having to bear the cost of huge back-end IT systems, according to Butler.

“It is one thing to connect things. It is another to transform those connections into something that really creates value for everyone in this whole ecosystem,” he said.

Addressing security and privacy
Integration is only one part of the IoT problem. The No. 1 challenge for businesses is cybersecurity, according to Gartner’s Lheureux. We have already seen examples on the web of things being hacked. Just recently, Cloudflare faced DDoS attacks on its servers from a simple IoT-type device.

“These things aren’t without consequences,” said EVRYTHNG’s Guinard. “Bringing things to the Internet is great because suddenly all things are integrated into the large web ecosystem, and you can do things we would never have thought would be possible. But the consequences are that these things are part of the Internet, so they become just as vulnerable as any server on the web.”

Still, Guinard believes the benefits of IoT outweigh the security concerns. “Would you use your smartphone or computer today without the Internet?” he said. “Connecting your phone and computer to the Internet makes them vulnerable, but there is such a broad array of things you can do with your phone and computer because they are connected to the Internet that makes sense.”

However, Guinard notes that IoT can sometimes be more vulnerable than laptops and smartphones because people don’t realize something like a smart fridge is actually a connected computer, so they don’t pay that much attention to the security of it.

The first thing Guinard says developers need to understand is that 100% security protection doesn’t exist. The only way to achieve 100% security on a network is to unplug the network. There are, however, some practices they can keep in mind.

Developers have to ensure they are protecting their IoT devices just as they would a laptop or computer by applying the same patterns they have used on other computers, like regular updates, strong networks, and strong passwords. Then they should also be using secure protocols to communicate over the Internet and locally, as well as using security keys and security certificates. Devices should also have the ability to upgrade and get fixes as new security flaws are discovered, according to Guinard.

“You need a way moving forward to upgrade it and ensure it has the latest security standards,” he said. “There is a responsibility both from consumers [and] also from the brands to keep the device alive and secure for as long as possible throughout the life cycle of the product.”

This is also where an API-led approach can come in, according to MuleSoft’s Sarid. “An IoT integration can potentially introduce additional entry points for a bad actor to gain unauthorized access to a system or systems,” he said. With an API approach, “It is possible to introduce greater security and increased visibility into the flow of data. Rather than connecting things point-to-point, every asset is a managed API, making it discoverable without losing security and control.”

Then there are privacy concerns facing the world of IoT. For instance, IBM’s Gorman explains if someone hacks into a thermostat or light-bulb app on your phone, they can tell if you are home or not; or a department store tracking your phone can monitor where you are, how long you were standing in front of a display, what you bought, and then go out and send you ads and other offers based on that information.

“IoT will enable a lot of intrusiveness that we are going to have to watch out [for],” he said. “At the same time though, it also adds a lot more convenience to your lives.”

In order to protect users’ data, and to ensure devices are working properly, developers need to have encrypted links and encrypted data for not only data in motion, but also data at rest. Identity management is also another way to protect data, and to ensure the person who is wearing a fitness device, blood pressure monitor or some other device is really that person, and there is no man-in-the-middle attack going on, Gorman explained.

EVRYTHNG’s Guinard adds that there needs to be a better framework for users to understand what they are giving away and what they are getting from it, instead of just accepting everything every time a sharing prompt appears.

“It is a whole new paradigm where everyone is your enemy, and whatever device you have, everybody you meet is trying to break your device, destroy it, hack it or do something wrong with it,” said Gorman. “You have to have that mindset where you are almost completely paranoid that once this thing is connected, you will completely lose control of it, so how do you defend against it, how do you keep the thing in a safe mode, and how do you ensure that some attack from the outside isn’t going to someone complete control over the device?”

Adding voice commands to IoT
The Internet of Things brings along new opportunities and possibilities for businesses to take advantage of. For instance, things like thermostats and refrigerators have been in our lives for decades, but today they can do new things like tell us what is in our refrigerator, or automatically adjust the temperature without any human intervention. One area that enterprise rapid application development platform provider kintone is looking to get into is voice commands for businesses.

The company recently announced new capabilities to bring voice commands to enterprise applications. “The goal is always to make a more efficient process,” said Dave Landa COO of kintone. “It is being able to access the process [and] move the process forward anywhere at any time, which is most convenient for the employee and the businessperson. We envision a number of scenarios just like from a consumer standpoint where voice activation might just be the most productive, most efficient way to interact with the database and move a process forward.”

It is still the early days of voice recognition, and in order to really gain value from these commands, natural language is going to have to improve. However, Landa notes voice recognition has already been getting so much attention and effort from the consumer side that the business side will be able to benefit from all that work.

The initial integration was built using Amazon Alexa Voice Service and Amazon Echo, allowing the solutions to interact with IoT-enabled devices in areas like agriculture, retail, transportation, and energy. Going forward, kintone is looking to build a platform and a foundation around enabling businesses and developers to add this capability to all connected devices. The company is building out its API and keeping an eye of voice recognition solutions that are rolling out.

“Our goal is to make teams work better globally and improve productivity. We just see this voice interface as a way to accomplish that,” said Landa. “It is something in which we see the future. It is a direction that the consumers are moving toward in terms of interfacing and accessing data, moving things along and taking action.”

About Christina Cardoza

Christina Cardoza, formerly known as Christina Mulligan, is the Online & Social Media Editor of SD Times. She covers agile, DevOps, AI, machine learning, mixed reality and software security. Follow her on Twitter at @chriscatdoza!