iStock connected

Sensor hardware has become cheap enough to sell in a pound store, wireless networking technology is robust enough to allow far-flung devices to communicate reliably, and in-memory data handling has made it feasible to process large volumes of information in real time.

Thanks to advances in these three areas, we are now able to embed communications modules on thousands of “things”, connect them over wireless networks, and process the data from all the “things” fast enough to help operators make informed time-critical decisions. If you aren’t impressed yet, just remember it’s only the beginning.

To get an idea of how big Internet of Things (IoT) will get, consider what Ericsson recently said about how much capacity future networks will need. According to the Swedish telecommunications company, it’s quite possible that by 2020, 50 billion devices will be connected.

Why believe Ericsson, a network equipment provider with every reason in the world to convince their customers - network operators - that the opportunity is bigger than it really is? You might be surprised to know that the Ericsson figures are an order of magnitude smaller than IDC’s expectations. According to IDC, by the end of 2020, approximately 212 billion “things” will be connected.

No matter where you get your market estimates, it’s safe to say that a boat load of vehicles, utility meters, personal handhelds, and even small sensors will be connected in the coming years.

Why should CIOs care? After all, IT have traditionally had little to do with what happens in operations, which is where most of the IoT applications are expected to be.

CIOs should care because all those “things” will be capable of sending information to central servers in real time for immediate analysis and decision support. As vice president and Gartner Fellow Hung LeHong puts it, “Much of the value from the Internet of Things will come from the data, making Big Data analysis a cornerstone of the success of the Internet of Things and a clear reason for CIOs to be involved.”

Hung thinks that CIOs need to take the lead in evaluating information requirements, and in ensuring that the organisation does what it needs to collect and analyse information from things, people, places, and systems. The transformational CIO will look for opportunities to implement the Internet of Things to change business.

Common components of an IoT Solution

Many of the Internet of Things implementations involve the operations part of a business - departments that typically lie outside the realm of IT. But CIOs should do everything they can to get involved, since most end-to-end solutions include four essential functions that the IT department has the most expertise on. These functions are:

  • Device management: To allow operators to install, update, and configure software on the different sensors, and to allow operators to detect problems with the sensors themselves - and even troubleshoot remotely - a device management platform is needed. Device management functions might include automatic back up of sensor content. The platform should also include security features, such as over-the-air encryption, password enforcement, and device wipe in cases where sensors holding confidential data are at risk of getting lost.
  • Filtering and event handling: Every end-to-end solution involves filtering out data that’s not relevant to a given application. Operators need to be able to configure events and actions that should be taken on each event.
  • An analytics engine: To handle all the data coming from devices in any substantial IoT solution, the analytics engine has to be able to handle big data in real time. The platform should be capable of correlating data from different device types and different locations and present the big picture to applications.
  • Connectors to back-office applications: The solution should provide connectors to make it easy for applications to fetch data from enterprise systems and to insert data into the same back-office applications and enterprise databases.

When these functions are included in a solution, IT should lend support. Ambitious CIOs are more than happy to do so, because they see IoT as a way of extending their influence into parts of the organisation where they previously had no say.

Industry-specific applications of IoT

While there is a lot of commonality on the technical aspects of IoT - the way devices are managed, the way data is filtered and analysed, and the way solutions connect to legacy systems - so far, there isn’t much commonality between different business applications.

In manufacturing, some industrial equipment manufacturers are optimising their maintenance processes and maximising uptime by instrumenting machines with multiple sensors. The sensors provide the manufacturer, distributors, and customers with real time machine status, battery charge levels, histories of operator intervention, and status of individual parts. When fault conditions occur, the sensors send trouble diagnostic codes to central systems for analysis, and possibly to alert operators.

In healthcare, patients wear devices that monitor vital signs and communicate data to medical practitioners. Health care professionals can configure triggers and events that cause appropriate alerts to be sent to the right doctors. Such devices might monitor heart rate, blood pressure, and blood sugar levels.

In energy and utilities, providers use networks of sensors to monitor the state of power grids. Operators can spot trouble early to prevent power outages. The sensors themselves also have to be monitored, because any failure in the measurement devices could mean power issues go unnoticed.

Common IT challenges with Internet of Things

As the market for Internet of Things is still in its infancy, there are still quite a few challenges enterprises face when one sets out to deploy an IoT solution. Here are some of them, along with some remedies:

  • Rapidly evolving technology and the rise and fall of different industry players makes it difficult to keep up with change. Look to implement solutions that are flexible enough to work with a range of device types and operating systems.
  • Each sensor produces a large amount of data. Taking all devices together, the task of handing all the information is daunting. Applications have to filter most of the data to get to what they need. Look for solutions that allow users to configure thresholds and specify events that should generate alarms.
  • When the first application is developed, care must be taken to ensure a second - or third - application can be developed later on to use the same sensor network. Most IoT solutions are dedicated to a single application. Over time, as people get more comfortable with the technology and its benefits, needs will arise to have multiple applications use the same network of devices. Make sure your implementation doesn’t rule out reusing the same sensors for future applications.
  • Operators must be able to install software on devices remotely. They must also be able to perform remote updates and configuration. All this has to be done without taking sensors out of service. Operators must be able to monitor devices to make sure they are working properly; and the operators need to be able to perform remote troubleshooting when problems occur. When sensors need to be replaced, it must be possible to swap devices with minimal down time. Look for a solid device management platform as part of the solution.
  • Whenever data is transmitted over wireless networks, security measures must be put in place to ensure sensitive data is not read and that spoofing cannot cause false values to be taken as real measurements. If data is sensitive, make sure implement the appropriate security measures.
  • Data must be correlated from multiple sources, and in real time. Many IoT solutions need to connect with back-office applications both to retrieve information and to post updates. Make sure your solution provides the flexibility to exchange data with existing enterprise applications and databases.

As IoT matures, development platforms will allow IT departments to implement different applications that share the same hardware and communications backbones. Instead of thinking about when and where to connect devices, enterprises will think about what more they can do now that devices are already connected. Enterprises will be able to use their networks of “things” over long periods of time, and in ways they couldn’t imagine in the beginning.

What CIOs should do now

As the technology is still quite new, and the vendor market is still quite fragmented, there is a danger in implementing large IoT solutions too quickly. Unless IoT solves a burning problem for your organisation, it’s better to start out with a small implementation.

IT directors should play an active role in those first IoT implementations, even if those initial solutions are in parts of the organisation IT isn’t typically asked to support. CIOs should develop a vision of where IoT will be used in the future and think of ways of using it to transform the business.

Consider what Gartner analysts have to say. They advise CIOs to do three things now:

  • Start taking the lead in figuring out the information needs your organisations has from its own Internet of Things. Information analysis will drive the business cases, promising greater efficiency, reduced costs, and increased revenue.
  • Create a team to become the experts on Internet of Things. This team needs to build knowledge, skills, and partnerships to take your organisation into the future.
  • Make sure your big data efforts are aligned with your IoT strategies, as data analysis is the driving force behind IoT. Whatever you do with big data, it should include the information you intend to get from your network of “things.”