To protect business servicing machines, GE turns to the industrial Internet | Hacker News Reader

To protect business servicing machines, GE turns to the industrial Internet

technologyreview.com

38 pointscpeterso9 comments

For a brief period I also worked at GE-Energy on their Smart Grid. This company has way too much bureaucracy to be effective in this space. The number of hoops you had to jump through to get a basic development environment set up were absurd. Many of the situations I ran into came right out of Dilbert land.

GE would be much better served by partnering with other companies and focusing on making hardware and machinery.

The real reason why GE is turning to "industrial internet" is because European firms have started a similar initiative previously, called "Industry 4.0":

http://en.wikipedia.org/wiki/Industry_4.0

Of course, GE will try to establish its own standards, which are incompatible to those of their competitors (e.g. Siemens).

For a brief period I worked at GE-IP. The reason I did not continue was, I felt I could add more value to the customer by using open source software and was not allowed to use it.

I was working in the mining industry. Where deep vendor lock-in is seen as the only path to success. GE follows this model to the core. Before a major revolution there needs to be more open source industrial software.

I work in an industrial space similar to GEs space and we generate high volumes of data from our machines in a distributed network. The data is really valuable in operating and optimizing our systems as well as doing reliability analysis. Somehow we convinced our customers to pay for a separate network connection as part of the agreements and do the VPN security. We do not operate over the customer networks except in certain cases and then both sides are super paranoid and isolate each other as much as possible.

We did a lot of the software in-house with open source except were we could not. Every once and a while we talk with GE-IP, OSI, Wonderware, and other large vendors in the space but talks always fall through in the end because its just really expensive. Even with steep discounts these vendors cost several millions of dollars over 3-5 years more than what we currently do for <$500K a year (most of that cost is the vendor software where open source is not available).

Vendor lock-in is a terrible problem for us. I'm pretty sure we do not get the value we are paying for. In one case we are using an 8 year old version because the 3-4 versions released since had major compatibility or breaking changes that we found unacceptable and then they started putting in intrusive always-on DRM-like licensing and actually cause more downtime for us so we don't use it and I'm trying to find a practical way out.

I want to use open-source and I would like my company to support putting our work back into open-source but I don't see it happening. There are way too many fears of loss of IP and effort to get lawyers involved to support it so it probably won't happen. I agree the industrial networking space is ready for disruption one of these days but its not exactly web pages that anyone can do out of school so it does not get open-sourced much but I think its coming soon enough.

What industry is this? There seems to be a lot of spaces where a few smart people could make a lot of difference. And it sure beats doing locsomobile amafacegoosoft.

EDIT: I am genuinely asking about the industry because I want to find spaces to expand to.

SCADA systems basically is the area you are interested in. Real-time industrial equipment monitoring (and control). Typically these are thick clients but people are clamoring for mobile and thin-client access to real-time data as well as historical time-series data. There are some very serious security threats in this space which is why the gorillas have leverage (not that they are really secure though).

Time-series historians are expensive if you log a large amount of data. There are several attempts at this like OpenTSDB which is a historian on hbase and is where GE is trying to go. We looked at their offering briefly and its hard to say if hadoop is good enough for near real-time but it looks good for long-trend analysis I think.

Ahh SCADA systems. Definitely something that should get some serious attention.

I'm looking for legacy or underserved complete verticals I can bring some modern technology to. Right now, that's low finance.

A challenge for GE is that it doesn’t yet have access to most of the data its machines produce. Courtney says about five terabytes of data a day comes into GE. Facebook collects 100 times as much. According to Richard Soley, head of the Industrial Internet Consortium, a trade group GE created this year, industry has been hobbled by a “lack of Internet thinking.” A jet engine has hundreds of sensors. But measurements have been collected only at takeoff, at landing, and once midflight. GE’s aviation division only recently found ways to get all the flight data. “It sounds crazy, but people just didn’t think about it,” says Soley. “It’s like the Internet revolution has just not touched the industrial revolution.”

The privacy, industrial espinoge, and confidentiality issues will always be interesting in this area.

Indeed; while it might be more speculation that reality (the firm is not saying much, but see http://krebsonsecurity.com/2014/02/email-attack-on-vendor-se... for what it's worth), I remember reading that it was thought Target was compromised through the path they supplied to a refrigerator and freezer vendor, which they didn't properly segregate or otherwise manage.

There's very strong use cases for this, e.g. those responsible for a store would much prefer being told quasi-immediately that a unit has failed instead of finding it full of ruined food when they arrive in the morning.

Reply on news.ycombinator.com