Can industrial analytics succeed in the cloud?

Is the cloud ready for industrial and machine data? I have been asking experienced practitioners whether they see new technologies fitting into the plant floor software market. On the topic of cloud, the jury was out – responses came back from “We think this is important and are moving there.” to “Certain applications will go there.” to “I don’t really see that happening, with the speed and security concerns.”

Manufacturing with analytics in the cloud

Can manufacturing performance analytics work in the cloud? GE is committing to it with an industrial analytics platform. Photo courtesy GE.

Did all of that uncertainty change last week?

On August 5, the young GE Software group announced Predix, a cloud service specifically designed and built to support industrial data and analytics. This service makes sense to me. Industrial data and analytics are quite different from other applications. Having a special cloud platform to host this type of application – and on which to develop this type of application – could make a world of difference.

Of all of the industrial applications, analytics is the most logical one to host in the cloud. It is typically not real-time, so that aspect of cloud performance is not a major concern. The people who need analytics data are typically in a variety of roles and locations throughout a company, which increasingly means scattered across the world.

GE says there are eight major differentiators of Predix from other cloud services. They are exactly the issues that have tripped people up on moving industrial software into the cloud to date:

1. Asset connectivity – Predix includes connectivity-as-a-service, based on both partnerships with global telecommunication providers and proprietary technology. Getting connectivity to industrial devices takes expertise that GE has.

2. Scalability for machine data – Talk about big data! Industrial equipment often generates data at a really rapid pace, and in formats that are widely varied, from time-series readings to imagery to sensor tags. This speed and variety would tax most platforms for analytics, let alone analytics in the cloud. Again, GE has cred.

3. Security – The number one concern for cloud platforms typically is security of mission-critical operations data . GE talks about adaptive security solutions for both developers and operators. I don’t have much detail, but my view is that cloud-hosted systems are typically far more secure than on-premises systems, as many threats come from human interaction, often inadvertent. See #6 below for an added element of security.

4. Governance and compliance – Regulations are always a challenge, and GE says it has experience in over 60 areas. As a global company that plays in regulated industries such as aviation, energy, healthcare and transportation, GE is both a customer and provider. This should help to harden the system for compliance in many countries and industries.

5. Interoperability – Seamless interoperability with other cloud platforms is GE’s claim, and it is critical they can deliver on it.  As companies move more into cloud services for ease of updates, offloading management and downsizing IT infrastructure support teams, cloud applications will need to cooperate across disciplines.

6. Gated community – GE is setting up an industrial ecosystem community that is not open to the public nor to other types of data and analytics. No doubt it will make this a less attractive hacking target than public clouds, at least in the short term.

7. Developer insight – Predix is designed to provide developers a view into not only the application in the field, but also every actor connected to it. This ability to monitor and potentially continuously adjust to meet changing conditions gives developers a way to see what might need an upgrade and potentially to also improve operational effectiveness based on the analytics.

8. On-demand availability – As with many cloud services, GE is offering a pay-as-you-go approach to provide quick access and easy scalability.

The development of Predix is a follow-on to co-founding the Industrial Internet Consortium (IIC) in March 2014, along with AT&T, Cisco, IBM and Intel. Predix will help host data and analytics from the industrial internet.

Is the jury still out? Well, probably. This is a brand new offering, and most of the applications to date are in utilities, not in manufacturing and production. Yet GE has a pretty good track record of delivering on its visions. Even as they do so, there will likely still be production companies (and GE competitors) who are not ready to jump into the cloud.

My guess is this is the beginning of a big move to host at least production analytics in the cloud. Those who don’t at least try it out may well find themselves as defendants in a losing battle after a few years.

Tags: , , , , , , , , , , , , , , ,

No comments yet.

Leave a Reply