agsandrew - Fotolia
To create a foundation for the Internet of Things, manufacturers must combine multiple components, such as gateways and platforms, with an integrated collection of hardware and software. Industrial IoT implementations also require a third puzzle piece: a secure place to store high volumes of data and run the specialized analytics programs that serve up valuable insights, such as why a production line is churning out poor-quality widgets.
A growing number of vendors are offering Internet of Things cloud options and making the case that a public cloud is the best place for industrial users to run their IoT projects. "Analytics are the key to the Internet of Things, and the cloud offers a platform that's easy to scale and has predictable monthly costs," said Alfonso Velosa, vice president of research at Gartner.
"Companies can start in the cloud to explore the value of IoT without having to make infrastructure investments," said Peter Christy, research director at 451 Research. "After that, they may decide to stay with the cloud or invest in a proprietary system."
Expanding Internet of Things cloud choices
Planners interested in IoT clouds can choose among offerings by IT industry veterans like Microsoft, Amazon, and Salesforce. Alternatively, manufacturers may opt for public clouds tailored for industrial operations from Bosch Software Innovations, General Electric, Siemens Industry and others. "If your center of gravity is the manufacturing floor, you'll want somebody who won't come in and mess up your equipment," Velosa said. "That may be a reason to consider one of the industrial technology companies."
GE is scheduled to make its Predix cloud platform widely available to manufacturers early this year. "It's not a pure public cloud, it's more like a gated community," said Rich Carpenter, chief software architect for GE Digital. "We'll do background checks on people entering the community."
In 2015, the company piloted the platform within a subset of its own manufacturing plants, including ones producing products for the aviation, transportation, power generation and utility industries, Carpenter said.
In one trial, IoT gateways collected data from sensors on computerized numerical control equipment and funneled it to the cloud for analysis. Tracking production losses was one focus area. The cloud-based analytics compared theoretical output rates to actual yields to determine the impact of machine downtime or losses due to quality problems.
"When we see the theoretical and actual rates start to deviate, we know that there may be a mechanical problem forming in some particular equipment," Carpenter said. "That information is valuable for showing us what steps can be taken right away to address the issue and prevent catastrophic failures."
The Internet of Things cloud enabled the company to simultaneously roll out the technology across multiple manufacturing facilities, rather than installing large-scale data storage and analytics technology one plant at a time, he said. The cloud also meant the company could expand or contract computing resources as needed for each analysis project.
"When we bring the data into the cloud, we can spin up a thousand servers to run a certain set of analytics across the enterprise," Carpenter said. "And when we're done, we just spin them back down. The elastic compute infrastructure is valuable for running analytics at scale."
Set clear industrial IoT goals
Upfront planning is required before manufacturers can contract for Internet of Things cloud services, experts said. That means close collaboration among the IT department, shop floor directors, the engineering team and others, Velosa said.
"IoT projects are often driven by a business unit, but IT needs to be involved in the up-front in decision making process," he explained. "This includes participating in strategic planning by helping to determine the priority of projects and evaluating cloud service-level agreements. Also, the IT staff truly gets analytics and security, so those are areas where it can provide value to the enterprise."
The IT department can also help decide if an Internet of Things cloud service is even viable for a particular project. On-premises analytics may be a better option when decisions need to be made quickly, such as when monitoring the operating temperatures of production equipment to avoid overheating. "Other times, such as when you need big-data analytics or contextual machine learning to make sense of a large data set, cloud solutions are often a good choice."
Planning with business goals in mind will help managers determine the right role for cloud computing. "Once you determine what you're trying to achieve on a business level, look at all of the available infrastructure options, from a private data center to the public cloud, to enable IoT in the most effective way," Christy said.
Learn about challenges of industrial IoT
Understand the role of M2M technology
Read about a mobile IoT application