In-Depth

WIRED: The Internet of Things Is Coming. Plan Accordingly.

IoT is here, but isn't yet well developed. Before jumping in, consider carefully what's available now and what your goals are for the future.

The concept of the Internet of Things (IoT) is starting to gain a lot of momentum; and, as usual with a hot trend, you're seeing a great deal of hype and hyperbole intermixed with a few factual statements. You see articles, reports, papers, and presentations opining on both the wonders and horrors of this technology in just about every industry journal and event.

Nearly every research firm has published its own taxonomy showing how it sees the market break down into submarkets, revenue opportunities and, of course, discussions of how enterprises should go about the business of deploying this technology.

Winners and Losers
Suppliers are jumping in to offer their own products and services, and are hoping that they can set the industry standards and force others to play their game. The name of this game, by the way? "We win and you lose."

This rush to claim territory is extremely similar to what you've seen each and every time a new, interesting use of technology emerges. As usual, the technology is based on things you've seen before, but that won't stop suppliers and research firms from coming up with new and different buzzwords and catch phrases.

Historical Roots
If you focus a bit more tightly, as with previous industry trends, the roots of IoT have been in manufacturing for decades. Numerically controlled manufacturing tools have been chugging away on shop floors for quite some time. As computers got smaller, faster and less expensive, they were built into more and more devices and tools. As inexpensive forms of networking were developed, they, too, were adopted and built into devices.

Now we live in a world in which very powerful, but tiny, computers are available that can control systems and communicate with the world. They're cheap, too; many are available for less than $100, and in some cases, even less than $25. Raspberry Pi and computers housed in a USB stick are plentiful. And as networked devices get ever more powerful and less costly, developers have been coming up with new ways to use the technology.

Some of these "innovations" appear useful and are likely to be popular. Others appear iffy enough that they're likely to die early in life. (That, by the way, won't mean that the idea behind them will die; they just might reincarnate when better, faster, cheaper and more capable technology appears later.)

Security and Privacy
As with many other tools that connect people and things, you're also seeing concerns emerge about privacy and security. You've seen reports of televisions "spying" on consumers by listening to everything said in a room, and then forwarding that data to some unknown server out on a network.

The suppliers haven't explained where the recordings of voices heard in the room are being sent, nor do they explain how long that data is kept or how that data is being used. While having televisions respond to voice commands is an interesting extension of the basic functionality of a TV, do customers want everything they say sent to some unknown place in the network for analysis and later use? You've seen similar capabilities added to mobile phones, vehicles and household appliances.

Defining IoT
Let's stop for a moment and define a few terms. As with other emerging industry topics, there are many different definitions of what the "Internet of Things" means. It's very early in the adoption cycle for this technology.

Suppliers, obviously, are doing their best to paint their offerings in the area of connected devices as early examples of this trend. They're also doing their best to present themselves as uniquely qualified to lead the charge to a more connected future. With so many definitions, it's easy to understand why many in the industry are confused.

If you examine a collection of these definitions, several common threads emerge:

  • Devices of all sorts now include intelligence to support and guide their activities. New automobiles have dozens of "electronic control modules," or computers that allow the vendor to offer new, interesting features and make it possible for the vehicles to be more useful, more efficient and offer greater value to consumers.

    Glancing around a typical home, you also see similar changes to appliances, entertainment devices and even things like clocks. Manufacturing has also seen numerically controlled devices in nearly every market and industry; these devices are more and more functional, and some can operate autonomously a great deal of the time.

  • Networking capabilities are being added to these devices, offering the hope of real-time operational control and custom usage of the device. This makes it possible for operational data to be collected and analyzed, and for enterprises and individuals to learn more about what they're doing and how they're doing it. Devices can speak with one another, making it possible, for example, for a television show to follow an individual as they move from room to room.
  • Devices can be personalized and demonstrate different behavior to different individuals, very much like today's mobile phones. Apps that offer new capabilities can be sold separately from the devices themselves, allowing individuals to fine-tune their own environments.
  • Suppliers are looking forward to being able to track everyone and learn what users are purchasing and rejecting, making it possible to target advertising even more finely. Refrigerators will be able to track what food individuals buy, and order products when supplies are running low. Health monitors will examine exercise or sleep patterns and report to doctors (and possibly insurers, hospitals and suppliers of health care products).

If you consider what the industry is talking about today, it's clear that IoT is having an impact now, and also changing what can be imagined in the future, in numerous areas of personal and public life, including:

  • Personal productivity. We're constantly learning about new intelligent devices that suppliers hope we'll love and carry with us all the time. Smartphones and tablets have become standard accessories carried by nearly everyone. Vendors are looking for new ways to connect users to the network. Watches are a recent competitive battleground. Who knows what's next?
  • Process control in manufacturing. Manufacturers have long had to deploy different technology and processes to build their products. This makes customization of products difficult and costly. If the manufacturing line was more intelligent, it would be easily possible for one product to be built with different features than the next one coming down the line. If you add in 3-D printing to the mix, manufacturing lines could be drastically simplified and the longtime promise of just-in-time inventory to be finally realized.
  • Retail. Point-of-sale and consumer personal productivity devices are increasingly linked together in new ways, making it far easier to simplify stocking and selling products. In some cases, inventory can largely take care of itself. As supplies are consumed, the retail establishment can automatically order more.
  • Health care. More and more, patients can be "instrumented" by wearing special clothing or carrying devices that monitor, in real time, all vital statistics and instantly notify caregivers when something unusual is happening. Doctors, nurses, therapists and other support personnel can often instantly communicate with one another; in addition, detailed records can be kept without burdening health care professionals. It's getting easier for specialists to provide care regardless of their proximity to the patient.
  • Vehicles. Vehicles already have become moving networks. It may be possible for all of the operational characteristics of vehicles to be monitored in real time to improve efficiency, offer new capabilities and reduce or eliminate the possibility of minor or catastrophic failure. Vehicles also could operate autonomously based on their location, the weather, surrounding conditions and other parameters, making travel safer, more efficient and less costly.
  • Appliances. Appliances are starting to learn consumer preferences and adapt themselves to operate more in accord with individual requirements. This provides the ability to monitor consumption of supplies and automatically order replacements.

If we examine all the devices we use on a daily basis, it's obvious that intelligence can be added to many of them, making our lives more convenient in the process.

While the vision is lovely, how to implement these dreams is still far too complex and device- and application-specific.

Emerging Frameworks
Each supplier has its own view of what IoT really means and how the technology should be developed. Unifying standards are emerging slowly. There are many IT suppliers offering tools, development frameworks, communication and networking tools that are tuned to specific requirements. Suppliers and enterprises have to sift through all of these offerings to determine if they should use off-the-shelf technology or build their own.

Here are a few IoT development frameworks that have come to the attention of Kusnetzky Group analysts. This is by no means a comprehensive or exhaustive list, nor is it in any specific order.

  • Eclipse.org has started an IoT working group that, in the words of the organization, "fosters the creation of extensible services and frameworks that enable IoT applications on top of open APIs." Many suppliers and academic institutions are using Eclipse as a foundation for their own work.
  • Kaa is offering its own open source IoT middleware platform, and is hoping to capture the attention of many developing communities to build solutions in specific markets.
  • Many suppliers are starting with the embedded Web and Web App services they've built into intelligent devices. They're hoping that they'll be able to stake out a claim by enabling today's Web developers to build device-specific apps and easily deploy them.
  • Publish/Subscribe frameworks have been offered by IBM Corp., Qualcomm Technologies Inc. and others to address how IoT applications can be built and sent to remote systems.
  • Resin.IO recently offered a complete development and runtime environment for Linux-based devices. Docker, Git, Yocto and other open source technologies are used as part of the Resin.IO framework.
  • Apple Inc., Cisco Systems Inc., Microsoft, Samsung, and a host of others are offering frameworks for IoT development focused on their devices and software. While each offers interesting capabilities, developers wanting cross-platform solutions might be forced to build separate solutions using each of these technologies, or force endpoint users to select specific devices.

Standards
While this vision offers many interesting possibilities, there are few standards in place. Because of that, suppliers have had free reign to select processors, memory, storage, networking capabilities, OSes, development frameworks and databases.

Because they're building their own computing environments, interoperability and compatibility from device to device and from supplier to supplier is still questionable Efforts are underway to create standards for machine-to-machine (M2M) communication, message queuing and security.

A challenge for the emerging IoT field is that integrating into today's computing environments is going to be a complex, tedious task; today's computing environments are often built upon multiple hardware platforms, OSes, development tools, databases, storage and networking standards.

This means developers must consider how their products will fit into today's management, security and development environments. What human and development languages must they support? Whose development libraries must be used? What embedded databases should be selected? Will these devices be able to support virtualization technology? If so, what standards will be supported?

It's clear that IoT is well established in manufacturing, but it's just emerging in other markets and applications. Because the requirements of a numerically controlled manufacturing device are very different than what's likely to be seen in a home appliance or a vehicle, cross-platform, cross-application standards may be slow to emerge.

Enterprise Planning Advice
Standards and frameworks are emerging slowly and differently in each vertical market. Once they emerge, cross-market/cross-platform standards will emerge. Enterprises, however, are unlikely to wait. Many have already begun planning, and in some cases, proof-of-concept projects are well underway.

Here are a few rules of thumb for your enterprise planning:

  • First and most important, the enterprise must determine what it's trying to accomplish before selecting a set of tools. Just because you have a hammer doesn't mean that everything is a nail.
  • Take the time to determine how IoT data will enhance enterprise operations, profitability and so on. If it isn't going to improve the enterprise in some way, it might be better to wait while standards emerge. Don't forget to consider how IoT data will enhance Big Data projects currently underway.
  • Don't forget privacy and security regulations and requirements when thinking about customer data collection projects.
  • Here's a big one: How will IoT deployments improve customer and staff experience? If jumping into IoT means making customer experience poor, the project is unlikely to succeed. Making customers do the hokey pokey to accomplish their goals isn't what it's all about. Jumping on the newest technological trend might be fun for the staff, but if it makes life harder for customers, they'll just go somewhere else.

Flash-Flood Warning
While the vision of IoT is enticing, it would be wise for enterprises to take their time to decide what they want to do, how it will improve the enterprise and how customers will benefit before leaping into the river. IoT has the potential of taking the enterprise where it wants to go, or just causing it to get all washed up.

Featured

Subscribe on YouTube