IoT solutions director highlights role of edge computing
One day after Dell said it would launch a new IoT division headed by former VMWare CTO Ray O’Farrell, the company’s IoT solutions director Kevin Terwilliger shared details about Dell’s IoT business. Speaking at the IoT6 Exchange Conference in Austin, Texas, Terwilliger said Dell’s $1 billion investment in the internet of things will be based on the information technology industry’s broadest infrastructure portfolio. Dell’s IT portfolio includes personal computers, servers and all the computers in between that will be increasingly important to IoT data processing. Terwilliger said many of these edge computers need to operate in difficult environments.
“Computing needs to happen in dirty, dank, dangerous environments where most PCs go to die,” he said. Oilfields, mines and warehouses are the types of places Dell sees demand for edge computing to process IoT data. Terwilliger said Dell wants to build IoT gateways that are “OT on the outside and IT on the inside,” meaning equipment that is rugged enough for operational technology but runs information technology software.
The convergence of OT networks and IT networks is seen as a building block of smart factory initiatives, and Terwilliger said this is not just a concept anymore. “We really are seeing these divisions coming together,” he said. The opportunity to realize operational efficiency drives companies to invest, he said, and the return on that investment can fund future IoT initiatives built around goals like improved customer experience, reduced business risk, and new revenue.
Dell’s IoT customers have included Emerson, Weir, Alps Manufacturing, and AeroFarms. Emerson and Weir use Dell’s solutions to monitor pumps and valves in locations that are hard for workers to access. Alps Manufacturing used sensors and data analysis for quality control on auto parts, and AeroFarms is exploring a wide range of use cases, Terwilliger said. For some customers, one of the biggest challenges is integration with existing infrastructure, Terwilliger said.
“It would be so easy if it were all greenfield,” he said. “A lot of times we are fitting into existing architecture. … We need to sit at the right place and bridge the right data back to the IT system.”
Sometimes Dell’s customer is not an enterprise but an equipment maker or systems integrator selling into the enterprise. Terwilliger said predictive maintenance solutions are more interesting to equipment makers than to the end users of factory equipment.
Clearing the fog
The “fog” is a catchall phrase for computers that blend some of the capabilities of cloud computers with the proximity and fast response time of edge-based gateways. Terwilliger said the fog could be a miniature data center on a factory floor, or a small data center at the base of a cell tower. Edge-based nodes may send data to the fog and then some or all of that data may go to the cloud for further processing and analysis.
Terwilliger compared edge-based computers to the arms of an octopus, a fish whose brain is distributed throughout its body. He said 65% of an octopus brain is in its tentacles, implying that a similar proportion of network intelligence might eventually reside in edge-based gateways.
“The cloud is the coach and the athlete is the edge,” Terwilliger said. “The athlete has to make split-second decisions but the coach is always teaching him. … The edge does machine learning and cloud does deep learning.”
