Your browser is not supported. Please update it.

13 October 2017

GE makes Predix pivot, exits data center biz, signs SAS and AWS deals

General Electric has announced two recent deals that caught our eye, suggesting a pretty major shake-up at its Predix Platform-as-a-Service (PaaS) inside GE Digital. Initially thinking that we were seeing the beginning of the end for Predix, a quick check uncovered a recent pivot in the Predix strategy – GE is no longer investing in owning its own data centers for Predix.

Naming AWS its preferred cloud provider was the source of our initial confusion, but signing SAS up to provide its Event Stream Processing for GE Transportation’s locomotives added to the questions – as the task at hand (data analytics) falls well within Predix’s remit. The AWS announcement focuses on GE’s IT applications, with GE migrating thousands of core applications to AWS, whereas the SAS deal focuses on expanding Predix’s capabilities via a partner.

GE did begin investing in data centers to support Predix in 2016, but has recently decided to use public cloud compute providers for their Infrastructure-as-a-Service offerings – renting the servers on top of which it runs the Predix PaaS that it provides to customers. However, it has since pivoted, evidently spotting a cost benefit in the new approach. Predix aims to provide the platform on which industrial applications are built, provided as a service, and handling the critical functions.

Mike Scott, GE Digital Europe’s Communications Director, explained that the public cloud providers made massive investments into their own infrastructure in 2016, when GE began to build out its own data centers. As GE puts it, Predix is the PaaS layer for the Industrial Internet, and the best place to build applications for the industrial domain, according to Scott.

The company plans on adding Microsoft’s Azure to the list of supported IaaS platforms soon, but others could well be in the pipeline – although GE is staying mum on the point, for now. Google would be an obvious one, and while their rivals like Baidu or Alibaba would make sense for China, GE Digital has already partnered with China Telecom for local Chinese infrastructure. Deutsche Telekom’s T-Systems already has a strong interest in providing IIoT services, and is another potential partner.

“Adopting a cloud-first strategy with AWS is helping our IT teams get out of the business of building and running data centers and refocus our resources on innovation as we undergo one of the largest and most important transformations in GE’s history,” said Chris Drumgoole, CTO and Corporate VP at GE. “We chose AWS as the preferred cloud provider for GE because AWS’s industry leading cloud services have allowed us to push the boundaries, think big, and deliver better outcomes for GE.”

As for the Amazon perspective, Mike Clayville, VP Worldwide Commercial Sales at AWS, said “enterprises across industries are migrating to AWS in droves, and in the process are discovering the wealth of new opportunities that open up when they have the most comprehensive menu of cloud capabilities at their fingertips. GE has been at the forefront of cloud adoption, and we’ve been impressed with the pace, scope, and innovative approach they’ve taken in their journey to AWS. We are honored that GE has chosen AWS as their preferred cloud provider, and we’re looking forward to helping them as they continue their digital industrial transformation.”

Moving to the SAS deal, GE is integrating its Event Stream Processing software into Predix Machine and Predix Edge Manager, which is then deployed on both the trains and inside operations offices. With an initial rollout plan to cover over 1,000 North American trains, GE is hoping that the real-time data streaming will help it spot when a train is not running optimally – saving it huge amounts in fuel and maintenance bills over time.

For machinery as intense as locomotives, even fractions of a percent in potential efficiency savings can add up to create a rather large incentive to adopt IoT technologies. Edge-computing is also the flavor of the month, and GE Transportation has evidently been swayed by the promise of having an analytics platform on the device itself – a rather large device, admittedly.

“In the IoT era, data is the fuel and analytics is the engine that drives value and tangible outcomes,” said Randy Guard, SAS EVP and CMO. “GE Transportation sought to develop an intelligent network to increase the value of its investment in connected IoT assets. They understand the need for analytics close to the edge, directly on the locomotives. With SAS Event Stream Processing, GE Transportation will benefit from a unified analytics environment and the ability for data scientists to code in the programming language they prefer, to drive valuable customer outcomes faster.”

For the stack itself, Predix is running on an AWS deployment, and receiving data from EdgeLINC – an evolution of GoLINC, running on the edge device. SAS’ Event Stream Processing software is integrated into the EdgeLINC instance, to analyze ‘high velocity big data while it’s still in motion – before it is stored – so you can take immediate action on what’s relevant, and ignore what isn’t,’ according to the sales pitch.

SAS, an American software developer focused on analytics, and apparently the largest privately held software company in the world, traditionally sells an analytics suite (called SAS, for Statistical Analysis System) to most of the Fortune 500. SAS says it helps customers at some 83,000 sites to make better decisions faster – and of course, with the advent of the IoT, there is a whole lot more data for the likes of SAS to ingest and process.

Analytics is a very hot topic in the IoT, and one that requires some careful navigation as users begin to merge their IT departments with their Operational Technology (OT) wings. GE’s Predix offers asset performance management (APM) analytics services itself, but, Amazon AWS, Cisco, Dell, HPE, IBM, Intel,, Microsoft, Oracle, PTC, and SAP are all competing in this market – hoping to add an analytics package to your PaaS suite.

“GE Transportation’s locomotives are equipped with edge devices that manage hundreds of data elements per second, to optimize locomotive operation,” said Garret Fitzgerald, GM Transport Intelligence at GE Transportation. “We use SAS Event Stream Processing in our Predix-based EdgeLINC offering to enable our customers with analytical insights on their data in real time. Although we’re in the integration phase of this project, the software has been easy to implement and use.”