Delving into an enterprise IoT initiative? Read this first
From the business problem to the technology, here's what CIOs need to know to get started on an enterprise IoT initiative.
The Monsanto Co. wants to help farmers solve a major problem: How to feed the additional 2.3 billion people we will have on earth by 2050, according to projections from the Food and Agriculture Organization of the United Nations. Throw in shrinking farmland and an industry that's traditionally slow to market, and the problem, it appears, becomes a crucible.
Monsanto, the world's largest seed company, has turned to the Internet of Things (IoT) in search of potential solutions. For Monsanto IT, the push has required upending tried-and-true protocols for the uncertain and the experimental. And they're not alone. According to experts, CIOs who decide to delve into IoT will be exploring new territory that includes new platform choices, new concepts such as edge computing, and new relationships with vendors that look and feel more like partnerships.
But the old problems don't go away either. Among the most important IoT issues facing CIOs is how to forge better relationships with the business where, according to experts, IoT opportunities often begin to take shape. The same cultural, language and trust issues that CIOs have been negotiating for years come to the fore in IoT. That's because many business people see IoT as extremely strategic and, according to Gartner analyst Paul DeBeasi, they can be almost secretive about it. "Even when IT folks found the business problem and began to understand the business problem, a lot of times, the business just put up a wall, and pushed them back and didn't allow them to get engaged," DeBeasi said of IoT initiatives at the recent Gartner Catalyst event.
Still, CIOs should make inroads now rather than later, if only for the sanity of their IT departments. "The business needs to get you involved, because they're only going to go so far," DeBeasi said. "Eventually, it's going to be dumped in your lap, and, eventually, you're going to have to run it."
How IoT works
For the uninitiated, IoT is a network of physical objects that have unique identifiers capable of producing and transmitting data across the Web automatically. While the technology itself isn't new, components such as sensors that used to be prohibitively expensive are now more affordable than ever, resulting in a wave of "smart" products flooding the market -- everything from bathtubs to baby monitors.
Monsanto's enterprise IoT strategy started as a way to reduce inefficiencies in its supply chain, such as preventing seed loss. Seeds that experience heat stress, for example, are unlikely to germinate. By outfitting the semi-trucks that transport seed from fields to processing facilities with sensors that measure temperature and geolocation, Monsanto's IT department was able to build a virtual window into the transportation environment.
Doing so gave the business an advantage: "Now, with IoT, if our grain gets heat stressed, we can dynamically route it to cooling centers or route it to the front of a receiving line to get the grain processed," said Fred Hillebrandt, infrastructure architect at the agro-chemical and technology company in St. Louis.
Hillebrandt and Monsanto were held up as a model IoT example during the recent Gartner Catalyst event, where more than one analyst remarked at how quickly the IoT trend is moving. "You started to hear about it a little a few years ago, but all of a sudden, it's here," said Lyn Robison. If Gartner predictions of 25 billion sensors generating $1.9 trillion in global economic value by 2020 play out, interest in the IoT won't fade anytime soon.
Put another way: CIOs need to start thinking IoT strategy, which begins by understanding how IoT works. According to Gartner, there are three components of any IoT service: the edge, the platform and the user. Here's how the three components break down:
- The edge: Where data originates or is aggregated, pared down and, in some cases, analyzed, according to experts. At Akamai Technologies Inc. in Cambridge, Mass., for example, edge computing plays a crucial role in real-time operating systems. Rather than transmit every signal from a sensor to a centralized data warehouse and bog down the network, Akamai collects the data at an aggregation point close to the user and transmits in real time only the data points that require immediate attention. Aggregation points aren't always necessary, according to Gartner analyst Drue Reeves, but the more IoT devices a business is gathering information from, the more critical they become. "When you have thousands of turbines or thousands of washing machines, you're going to want some aggregation in the middle or you're going to overflow your IoT platform with data," he said during his presentation at Catalyst.
- The platform: Where data is ingested -- typically in the cloud, analytics are performed and an often internally developed algorithm takes an action, according to Gartner's DeBeasi. Incoming data is sent to a real-time stream processing engine, which decides if an action needs to be taken right away or if the data can be tucked away for future use, according to Reeves. The platform also performs big-picture analytics by, say, integrating historical data with real-time data to look at trending analysis; it also contains a policies engine and an orchestration engine to manage the platform, Reeves said.
- The user: Where data drives a business action. According to Reeves, data that's been analyzed can move from the IoT platform to a user in one of three ways: The user can use an API to "call" or query the data, the IoT platform can call out or signal to the business user when it finds a predetermined set of events, "or you can do both of those over an API bus," Reeves said.
An IoT strategy
The recommended steps to bootstrap an enterprise IoT initiative will be familiar territory to many CIOs. Gartner suggested they build a team, find a business opportunity, build a prototype and decide if the project is worthwhile enough to invest in. But several of the finer points are worth highlighting.
For example, DeBeasi recommends CIOs start by appointing what he calls an IoT architect -- someone who:
- understands IT and has a basic understanding of IoT technology;
- has a deep curiosity, as well as a willingness to learn and share; and
- understands operational technology and how IoT will impact business operations.
That last trait is what makes IoT so different, according to DeBeasi. "It's not enough to be technically savvy; this person must be business savvy as well," he said. "They understand the business process; they're a good communicator."
At Monsanto, Hillebrandt, who works for both the strategy and architecture organization as well as the supply chain organization, has taken on the role of IoT architect (not his official title). Last February, he was tasked with leading an IoT initiative for the company. "I'm still focused on supply chain, but I was asked to put together an enterprise strategy around IoT -- a platform strategy," he said. (For the record, Hillebrandt's first step was also people-focused: He put together a small team of influencers who were already tinkering with IoT or had expressed interest in aspects of it and had credibility with senior management, he said.)
IoT platform vs. point solution
Another aspect CIOs will want to focus on is the platform itself. When deploying, CIOs have a couple of options, according to DeBeasi: They can either select a point solution or a more general purpose solution.
How to bootstrap your enterprise IoT initiative
1. Build a core team. Gartner analyst Paul DeBeasi said to start with an IoT architect who has a keen understanding of IT, but also understands operational technology and how IoT will impact business operations. DeBeasi also recommended that the IoT architect partner with an IoT analyst, someone with an even deeper understanding of the business.
2. Discover the opportunity. DeBeasi used the word "discover" rather than "find" to convey how difficult of a challenge this might be for IT departments. Since IoT initiatives are starting in the business, figuring out how to build strong partnerships will be key.
3. Create a data model. DeBeasi suggested three steps when creating a data model: Define your data streams; define how your platform will process the data; and define the business action you want to take.
4. Build a prototype. Businesses will need to be developing and testing algorithms and automated responses, according to DeBeasi. "In the Monsanto example, this is where they have the action of prioritizing the seed coolant or the processing of the seed," he said.
5. Invest in the project. Verify the business value, assess the feasibility of the project and consider the scope, DeBeasi said.
Source: Gartner Inc.
IoT point solutions like BigBelly Solar trash compactors are a low, initial investment, can be deployed quickly and can provide rapid innovation, according to DeBeasi. "You're not defining any of the data, you're not building the sensors, you're not building the dashboard," he said. But there are drawbacks, such as the inability to integrate a point solution with a larger enterprise IoT initiative, thereby creating IoT silos, and that's the last thing CIOs want, according to Monsanto's Hillebrandt.
Fred Hillebrandtinfrastructure architect at Monsanto
He compared it to the 10 instances of Salesforce that Monsanto currently supports. "Every time we stand up a CRM for a region, we're building it all from scratch," Hillebrandt said. "That's not what we want to happen with IoT. We really want to have one platform, because that's going to get us the agility -- the innovation through sharing the data to build these higher-value applications."
To avoid silos, CIOs can either build an IoT platform internally, like Monsanto did, or follow Gartner's recommendation of deploying a "universal IoT platform" from Amazon Web Services, IBM Bluemix, Microsoft Azure or Google Cloud Platform. By getting all lines of business on the same platform, IT departments "can begin to then do federation or integration of information," DeBeasi said, as well as build additional capabilities into the platform -- another characteristic point solutions lack.
Selecting a universal IoT platform over a point solution won't happen without the business. And, unlike Monsanto's case, IoT opportunities won't normally start in the IT department, according to experts. (And even in Monsanto's case, one of Hillebrandt's goals is to build a self-service platform for the business. "It's our intent to deliver business value by deploying an IoT platform that lets the business chase after cost optimization and bring new digital products to market," he said.)
For most CIOs, the process will reflect something closer to what Jonathan Reichental, CIO for the City of Palo Alto, Calif., is experiencing. IoT initiatives, such as deploying a smart grid or implementing smart traffic signals, are coming out of different departments -- utilities or traffic -- in a fragmented way. But unlike what Gartner analysts observed at other businesses, Palo Alto departments aren't hiding IoT initiatives from Reichental. "I'm pulled in occasionally for discussions with vendors," he said. "Or, when we have to issue an RFP, my advice is sought."
The relationship between IT and city departments operates like this for a couple of reasons. Technology these days almost demands it, Reichental said. "In today's hyper-networked environment, no department can really go rogue without having to plug into the network, and get ports open, and conform to things like backup strategy and archiving," he said.
More pragmatically, Palo Alto has adopted a governance process, supported by the city manager, on steps that need to be taken before a technology project moves forward. Reichental described it as a funnel. "It goes through and we evaluate it for architecture, security, data backup, staffing, budget, and what capacity and skills we have," he said. In some cases, projects are deemed "not a good idea," and that conclusion is supported by the data gathered.
In the case of IoT initiatives, Reichental sees the practicality that a project like digitized parking meters could have for Palo Alto. By this time next year, he hopes to have a more centralized IoT vision for the city, but in the meantime, he doesn't want department efforts to live in the shadow. "I'm fully supportive of a partnership between a centralized IT and the needs of the departments," he said.
Besides, he said, if the relationship between business and IT isn't strong, the work doesn't get done. "You've got to have really good credibility and trust," he said. "And then you've got to have a governance process in place, endorsed by the CEO or the city manager downwards."
Building trust is easier said than done, so here's a basic piece of advice Gartner analysts provided to get started: Scrub the term IoT out of existence when talking to the business.
"Business units won't call it IoT," Reeves said. "They'll call it something else, like digital business or the digital oil field or the connected car." In fact, scrub the tech talk out altogether. "Learn the business language," DeBeasi said. "Learn what's happening in the business; learn how to communicate."
Let us know what you think of the story; email Nicole Laskowski, senior news writer, or find her on Twitter @TT_Nicole.
How Jonathan Reichental is trying to help build a smart city
Manufacturers look to IoT gateways for innovation
Internet of things: Food safety apps poised to flourish
Grow IoT apps using a cloud model to fulfill future business needs