Cloud Computing: 8 Hidden Costs

GE will enter the cloud services market early next year, offering a Platform-as-a-Service (PaaS) designed to capture, store, and analyze industrial-scale machine data.

In GE's view of the future, the consumer Internet will be dwarfed by a second phase, machine-data-based Internet that tracks and monitors civilization's engines and systems of all types, keeping them running. The GE Predix Cloud will, as the name suggests, offer GE's Predix software system for capturing data out of large manufacturing or industrial operations and perform analytics on it.

GE currently uses the Predix analytics system to monitor and maintain its large-scale industrial products, such as wind turbines, jet engines, and hydroelectric turbine systems. Predix was developed by GE Software, a business unit of the company.

With the Predix Cloud, GE is trying to make Predix a scalable service offered on cloud infrastructure. GE will move its own systems onto the Predix Cloud in fourth quarter 2015, then make the Predix Cloud available in 2016 for machine data users. GE spokesmen said there are likely to be many users of a machine data analytics service, if it were convenient to tap into it and use to capture and store data. For example, Pitney Bowes, a user of the Predix analytics system, is an early, beta user of the Predix Cloud to extract and analyze data from its postage metering machinery.

CEO Jeffrey Immelt said in an Aug. 5 announcement about Predix Cloud that the service will keep its focus on machine-generated data, and will be better at capturing it than more general-purpose service providers. By using Predix services, "A more digital manufacturing plant means more products are made faster," he said.

[Want to learn more about GE's Predix software? See GE Prepares For Industrial Internet With Cloud Foundry.]

Uses for a Predix Cloud aren't limited to ones that reflect GE's interest in manufacturing large-scale industrial products. For example, the Predix Cloud could capture data flowing out of a hospital's patient, imaging, and treatment systems. "A more digital hospital means better, faster healthcare," Immelt added. Likewise, an oil exploration company can use Predix to achieve "better asset management and more productivity at every well," Immelt said.

Essentially, Predix is where big machine-generated data meets analytics, with every company applying its maintenance expertise through software systems that have access to the data. "A cloud built exclusively to capture and analyze machine data will make unforeseen problems and missed opportunities increasingly a complication of the past," said Harel Kodesh, VP and general manager of Predix, in the Aug. 5 announcement. Kodesh said Predix-based systems will allow companies to monitor their products as they work, detect anomalies and early indicators of pending failure, and fix them before they break.  

Predix "will unlock an industrial app economy that delivers more value to machines, fleets and factories," Kodesh claimed. Such an economy will need armies of programmers who understand the maintenance of modern systems and create applications that use the data flowing out of them to ensure their own maintenance.

A cloud service supporting use of machine data will need to be highly secure and ensure that the data stored there has the protections needed to keep it in compliance, such as electricity supply data subject to supervision by a public utilities commission. With GE's experience, a Predix Cloud would be able to supply tools for asset connectivity, support for formatting machine data, and "industrial-grade" security.

It's conceivable that complex households with many interlocking digital systems will one day need an Internet of Things-type data center to capture and process data for optimum operations as well. But, in general, Kodesh was skeptical that small-scale users will play much of a role in a Predix Cloud, At the Cloud Foundry Summit in Santa Clara, Calif., last April, he told InformationWeek his appearance at a consumer conference would be equivalent "to a crop duster showing up at an organic gardener's convention."

"GE knows industrial machines and related data analytics better than anyone," said Roger Piloc, chief innovation officer at Pitney Bowes, in the Aug. 5 announcement. The company's postage metering division runs asset performance management software on Predix to keep an eye on its metering equipment.  

GE believes the industrial Internet will generate data twice as quickly as the consumer Internet and require an investment of $60 trillion over the next 15 years. To cope with the data, GE is building out its Predix system using the Cloud Foundry open source platform. Predix can run on any cloud infrastructure and currently runs on Pivotal Software infrastructure. Customers will have the option of using Predix services on the Predix Cloud, or using a Predix system on some other infrastructure.