From wearable devices to computer sensors that monitor the manufacture of products and water accumulation for farmers, the era of the Internet of Things is here. Consumers want their FitBit, Apple Watch or other device to capture the most accurate and timely results, but effectively deploying enterprise IoT devices in business is a whole other matter.
“The gadgets you buy at Best Buy are not the same as what you need for manufacturing or the production line. You have to worry about security and manageability. It’s a different set of capabilities,” Jack Gold, principal analyst with J.Gold Associates, told Datamation.
And while companies manage laptops, tablets and even smartphones from their existing infrastructure, Gold notes most don’t have systems in place ready to manage IoT. “What do you do when a device that has a 32 cent CPU doesn’t know anything about your VPN or you need it to talk to Salesforce?” Gold asks rhetorically. “Yes, there are apps designed for this in some cases, but there are also security issues. How do you make sure these devices are secure?”
Decisions also have to be made about how the data is going to be analyzed. “If you can’t get actionable intelligence, why deploy IoT at all? You have to be able to do analytics in real time and get the data you need to run your business,” says Gold.
Adam Wray, CEO of Basho Technologies that offers a distributed NoSQL database, has many customers using the software for a variety of IoT deployments. He says companies considering an IoT deployment face three major challenges.
The first is security. Personal data transmitted by devices that monitor health must be secure to protect personal privacy. Business and industrial data collection needs security for competitive reasons, compliance and simply to prevent its misuse by unauthorized entities, be they competitors or even someone in your own organization who is not set up to manage the data correctly.
The second challenge is managing the enormous amounts of data IoT devices are capable of generating today – with even more in the future. Wray notes that Tesla generated but a few data points when the luxury electric cars were first released, but that quickly grew to hundreds.
“There is a full expectation from what we hear from our partners that there will be thousands of endpoint sensors in cars and instead of passively gathering data every few hours, millions of cars will be collecting and sharing data every few seconds,” says Wray.“That’s big data no one has ever seen before. It’s going to put massive stress on the network and require a lot of processing power.”
Third is latency. In many scenarios the real value of IoT comes from the real time collection of data. “You want to be able to make the data you’re collecting actionable,” says Wray. One of Basho’s customers, The Weather Company, collects data from millions of endpoints. But as Wray notes: "It's physically impossible to effectively transmit terabytes of data to a central location for processing.” Instead Basho’s technology is able to identify what parts of the huge data stream is useful or actionable and only brings that in for processing.
“There are limits to what you can do,” says Wray. “The speed of light is the speed of light.”
Leveraging Transactional, time-stamped data
Bryan Kester, who heads the IoT group at Autodesk, is in solid agreement that scaling is key to a successful IoT deployment.
“Anyone can collect data at a small scale. If you have a hundred to a few thousand products and some off the shelf tech, you can put together a solution in a few months and operate it in the data center or the cloud,” Kester tells Datamation. “But what will happen as you as add more is that the data management grows and most people don’t know how to design distributed systems that can deal with a lot of time series data.”
Autodesk offers a cloud-based service focused on transactional, time-stamped data spread across tens of thousands of its servers. “It’s the power of the cloud that lets you offer a cheap cost to do lots of number crunching you couldn’t do before and an application server on top,” says Kester. He notes that even presenting all that data effectively requires the kind of specialized system Autodesk has developed. “A million data points on a chart would stop most web browsers from working.”
When is it time to jump in?
Analyst Gold concedes technology is often hyped well ahead of when most companies can truly take advantage of it. This is true either because companies don’t have the systems, infrastructure or people in place to leverage it, or the technology itself hasn’t been fully developed.
In the case of IoT, Gold says it’s still very much early days, but the technology has proven itself to be a game changer that warrants, at a minimum, test deployments and experimentation.
“If you don’t, you can bet your competitors are already moving forward and you’ll be left behind,” says Gold. “Look at folks who took too long for client server or didn’t embrace cloud computing for years. They’re at a competitive disadvantage. I’m not saying to just throw money at IoT. You should have a strategy and get started. Big companies generally take three-to-five years to develop a strategy around deploying new technology. The time to start architecting is now.”
Already he's seeing many examples of how IoT is or can potentially save companies big money. Gold points, for example, to a device called BlackBerry Radar for the trucking industry, from smartphone maker BlackBerry. Radar could help pharmaceutical companies track pallets of drugs that risk being ruined in transport due to overheating or other environmental conditions. Thanks to 4G and newer 5G connectivity, Radar can offer real-time monitoring of the temperature and send alerts when there's a problem.
"We're talking about tens of millions of dollars or more in potential savings to prevent spoilage or other damage," says Gold.