Big hurdles? You bet. But let me offer a few little lessons…
1. You’re going to generate mountains of data
Machines put the “big” in big data. A smart mining site can generate two petabytes per day. Vibration-analysis systems can soak up 200,000 signals per second. Some utilities monitor more than 27 million end-points. Harvesting and exploiting this data is the key to becoming more efficient, reducing energy and moving toward true sustainability.
2. Variety (in standards) is critical
Standards proliferate in operational technology (OT). A wind farm, for instance, might actively capture data from 300,000 data sources (“tags”) on a continual basis, served up in 140 different formats. Different markets have their preferences: OPC and OPC UA are popular with manufacturers. BACnet, meanwhile, is popular in the building-management market. Why such variety? Many of these devices are placed in extremely challenging environments and must last for years without human intervention. In the U.S., for instance, the average age of an electrical transformer is more than 40 years old. Performance trumps convenience.
3. The cloud isn’t the solution to every problem
Today’s cloud systems are magnificent. You can spin up hundreds of thousands of servers in an instant. But they aren’t perfect for everything. Clouds are often located miles from your devices and your data. Clouds introduce latency and the risk of communications failures. Take it from your engineers in operations: a substantial portion of your data will be stored and analyzed right where it was born.
4. This technology must be as easy to use as a faucet
Think of the power grid or the water system where you live. These are incredibly complex networks that are, at the same time, incredibly easy to use. You can attach a wide variety of devices—lightbulbs, household appliances, fire hydrants, faucets— that are all supposed to work perfectly at a moment’s notice. Water and power networks are so reliable, in fact, that when they break down it is headline news. This same infrastructure approach—ease of use, reliability, seamless scalability, broad compatibility—will be absolutely essential in IoT. Customers, service providers and the original manufacturer will access smart devices with very little training. Different types of hardware will be plugged into networks and removed without edicts being handed down by IT. Your employees and customers are going to want to run analytics without waiting for data scientists to scrub the data. Like water when you turn the nob at your sink, the data just has to be there.
5. Take it easy, fella
Don’t boil the ocean all at once. A lot of companies tend to start using IoT for predictive maintenance. Dong Energy uses IoT to reduce the frequency that technicians must boat out to inspect offshore turbines. The company anticipates saving 20 million Euros a year by 2020. Once you can show the ROI, everyone from the CEO on down gets more comfortable with the idea.
6. IoT is a revenue-generator
When people think of IoT, they often think of cost savings. The IoT can reduce energy, cut emissions, etc. But it’s also going to serve as a revenue-generator…the basis for selling new services such as predictive maintenance. Every product will have an attached service (consider flow-serve).
7. Big benefits
After the struggle and pain of adoption, there is reward. We’re already seeing success stories in likely (and unexpected) places. Irish distiller Pernod-Richard has launched a plan to cut energy consumption by 50% while doubling production at some of its distilleries, simply by using data better. The San Diego Padres, meanwhile, are on track to cut resources by more than 25% over the next five years by smartly monitoring energy usage.
Cheers to that…let’s play ball!
Martin Otterson is senior vice president of customer success with OSIsoft.