Don't forget the "edges" of your industrial big data picture

GE Automation

Don't forget the "edges" of your industrial big data picture

Posted in General on Wednesday, 07 September 2016
Don't forget the "edges" of your industrial big data picture

Remember when Linux and OpenOffice were going to liberate us all from paying for software? The reasoning was that we had collectively become technology-savvy enough that we’d prefer complex, powerful, open-source tools that required a bit of technical knowledge to use, but would relieve us of the burden of enriching evil software companies. Well, these days it feels like all that’s old is new again. The media is abuzz with stories hyping the various free or low-cost entry ramps to Internet of Things (IoT) and Big Data applications – including free/nearly free edge devices! The emotional discounting of technology now even extends to hardware. Ouch! For a firm that sells stuff, hearing voice after voice declaring that customers shouldn’t expect to pay for stuff can really wear you down

Interestingly, the voices we hear generally aren’t coming from customers. Media outlets, tech bloggers, even politicians (who simultaneously extoll their delight at “free” tech, while insisting that all schoolkids should be en route to careers as engineers and developers) can be heard chorusing for free stuff. But our customers aren’t. We like free stuff as much as the next guy, so thought it’d be good to see whether we have overly tolerant, rich, or otherwise odd customers. Some common threads emerged as we investigated:

  • In a world of chronically “lean” (not to say understaffed) technical teams, there’s not much leverage in a DIY kit. One way of looking at a packaged product is to think of it as the embodiment of the requirements and experiences of every customer that bought it before you… And that applies both to basics like tools and methods for configuration and deployment, to whatever in-the-box functions or outputs are part of the package. Free stuff tends to be – “components”…
  • The skill sets needed to develop software building blocks aren’t the same as those needed for getting value out the other end. And our customers seem to prefer enhancing process or asset engineering skills to building up foundational programming capabilities
  • Ditto for hardware – Sure a Raspberry Pi is cheap. But there’s more to making an edge device suitable for different industrial applications than a bit tougher box and mounting for the board and flash card. From physical and firmware design that recognises that many industrial assets use older comms topologies, connections and protocols, to the value placed on vendor-led compatibility and standards testing, our customers valued the “pre-work” built into the sorts of edge devices we put forward
  • Other interests? Oh yeah - security, security and security. Ensuring that sensitive data isn’t compromised is one aspect. So is ensuring edge devices don’t compromise the operation of a key asset or process. Something like certification to Achilles standards is another aspect of work/time/cost that customers prefer we vendors do
  • Every Industrial IoT application calls for a Swiss Army Knife – the final thought that comes up repeatedly was that there often isn’t an easy way to “jam in” black box after black box as asset manufacturers, owner/operators or services organisations look for new capabilities for data collection, or delivering apps back to the field. So being able to put things like an Industrial Internet-ready controller, or an HMI panel that’s also a terminal service client, and has a big data collector built into it become serious advantages for “not free” stuff…

We think there are some broader issues coming out of this: there’s a gap in much of the talk about Industrial Big Data and analytics. We think the gap is there because so much of the conversation is dominated by voices concerned with value propositions and “the IT end of things” – reporting and analytics. And the proliferation of exciting sounding firms that can do things with data generates its own buzz. But to listen to our customers, more time needs to be spent offering practical tools and guidance to lay the foundations that make data accessible – with particular focus on industrial-system-ready edge devices.  Time to get more voices talking about why connectivity for the Industrial Internet is very different than for consumer IoT…


There are currently no comments associated with this blog post.