Those of you who are still with me after all these months are aware that I believe that there is a coming revolution in data. I reviewed Chris Stakutis and John Webster's book, Inescapable Data and found kindred spirits there, and was again pleased to find in Glenn Almendinger of Harbor Research another, whose take on the coming data deluge he calls "Ubiquitous Data." Same difference.
The Zigbee Alliance open house today, in Schaumberg IL (boy it is nice to go to an important meeting so local that it was a five minute drive from the office), was not about technology as much as it was about data.
We have a technology, called mesh networking, that we know works extremely well. Our two biggest networks are not only based on the mesh networking topology, they also interoperate on it. Those are the Internet and VoIP telephony, and the land-line based telephone switching network itself.
We know mesh networking is robust and viable...it was designed to survive nuclear war, for pete's sake.
So what do we have to do to make it robust enough to make it in the industrial automation space?
First, we have to consider whether the Zigbee folks really know what they are talking about, marketing-wise. Most of them have fixated on the home automation market...why, I don't know. I am an extreme skeptic about the adoption rates they are talking about in home automation, beyond the 'early adopters.'
I submit Mrs. Boyes as the typical example. This is a woman who has a teaching credential, nearly enough credits for a master's degree in special education, and who is flatly brilliant. Her response to me when I said I could get a Zigbee network up and running in the house so that we could turn the lights on from anywhere and do all the other stuff people are talking about was, "What the heck do I want that for? Are you too lazy to get up and turn on the light?"
In other words, there is no value proposition that will convince the Mrs. Boyeses of the world to adopt Zigbee-enabled mesh networks at anywhere near the rates envisioned by the proponents.
HOWEVER...there is a value proposition, in fact more than one, that will convince industrial automation segments to be much more mainstream than simply early adopters. Industrial automation is the only place we can get more revenue from our already pressed manufacturing plants. We need to be able to provide continuous uptime, prevent incredibly expensive unscheduled downtime, and predict and prevent dangerous accidents like the "coker launch" in Anacortes, WA a few years ago, and the repeated accidents during turnover and startup at Phillips and BP facilities in Texas, among others.
We need to be able to connect these sources of data to our enterprise networks desperately, especially since SAP and Microsoft have finally figured out that real-time process optimization means real-time control, and not simply a jazzy automated general ledger.
There were some very cool things on display, from people like Ember, who are devoted to making stuff for OEM use, mostly, and for home automation. But the coolest stuff, by far, for this process automation junkie, was the stuff that Sensicast, Cirronet, Freescale, and others had on display that really did have existing applications.
As y'all know, I'm not shy, and I buttonholed Bob Heile, the Chairman of the Zigbee Alliance, and told him what I thought was the next big push for Zigbee...and he didn't disagree.
One of the things I talked about a lot at the meeting was the article in the current CONTROL on page 103 "Users Want an Industrial Wireless Standard."
In that article, I noted that Zigbee scores poorly on Interoperability. I need to make a small amendment to this statement. According to Bob Heile, the Zigbee standard is written to permit what I call "incompatible interoperability," by which I take him to mean that multiple proprietary applications can interoperate on the same mesh network using the Zigbee protocol. This is a very interesting take on what may be the standard-killer--- just like interoperability was the standard killer for industrial fieldbus. In other words, it doesn't matter what you run in the application layer, if it is Zigbee standard compliant. You can run competing applications at the same time, over the same network.
This is a very interesting development.
I'm going to the Wireless Sensors Solutions Conference (in Chicago, Sept. 26-28) to do more wireless-scouting.
Tell me what YOU think.
Walt