Why H.A. Manufacturers Fail: Revisited

Some folks recently commented an old editorial of mine called Why H.A. Manufacturers Fail. It appeared in the March 2006 issue of CE Pro, and was written in January of that year.

All respondents raised very good points, so I would like to bring the discussion up to the present. I very much appreciate the feedback and welcome a healthy discussion here in 2007.

Not There Yet

Eloy Paris commented in the blog:

I disagree with the "All IP" point. These days everything is networked, and the predominant protocol is IP.

Fair enough observation, but my IP comments have to be understood in their context -- the date, the audience and the product category.

Eloy goes on to say (edited for punctuation):

You say, “Folks from the PC industry erroneously assume that TCP/IP is the best and only standard for whole-house communications and control." My experience has actually been very contrary to this. I’ve seen support for lots of different things (X10, Z-Wave, etc.) but not TCP/IP. I am actually looking for TCP/IP-based solutions since my house is completely wired for TCP/IP on top of Ethernet

Well, that was pretty much the point back in January 2006 – that most things in the home-control world were not native TCP/IP communicators. From your experience today, you will see that that is still the case. However, IT pros had erroneously assumed that everything in the home control world is IP. Manufacturers came into the home-automation business thinking, "I can create the most advanced IP networks in commercial spaces, and I'll just apply the same techniques to the home."

They failed.

For data and some A/V applications, no problem (well, fewer problems). For home automation vendors, big problem.

I bring up IT folks not to belittle their obvious contributions to the home-controls industry, but to compare them to the automation and A/V pros who knew all about challenges in the home. IT experts are arguably the most important asset coming into this market.

The Nature of Home Automation

Back in 2006, virtually all of the intelligent subsystems -- security, audio (whole-house and theater), video (whole-house and theater), lighting controls, HVAC systems, motorized shades, you name it, were controlled either via RS-232 or relays.

With precious few exceptions, they still are today. Sure, most subsystems now have Ethernet ports on them, and they can be controlled via the Internet (which many people mistake for IP-controllable).

Let's take the pre-2006 case, when they were virtually all RS-232 or relay controllable. For an integrated IP-based system, each of those devices needed an RS-232-to-IP adapter, such as the ones made by Lantronix--which back then were about $250 each, if I recall.

Easy enough. But IT-oriented mfrs and dealers mistakenly believed that the physical bridge would do the trick. Way wrong. For every single subsystem -- Lutron Radio Ra, Lutron HomeWorks, Denon receiver, Marantz receiver, this security system, that security system, they needed to write the software drivers to enable IP communications.

If everyone agreed on a single brand and single model for their subsystems, it would be no problem. But there are a lot of drivers that must be painstakingly written and tested. Successful home-control companies today may have hundreds or tens of thousands in their databases

Today's IP

Today, as mentioned previously, most meaningful subsystems do have an Ethernet port on them. But that doesn't mean they can plug and play with third-party controllers and subsystems. There is no standard communications protocol for home control, so the drivers still need to be written.

You cannot plug any IP-controllable thermostat and any IP-controllable security system into a router and have them automatically discovered by a PC with third-party home automation software, and incorporated into a unified GUI.

You can have a pretty easy job if all of those devices have built-in Web servers that serve up GUIs (graphical user interfaces), but that makes for a very messy home control "interface" with every product having its own look and location.

So, bottom line, although not enough traditional home systems have two-way Ethernet for control and feedback, many more do today than did one year ago.

But that doesn't make them any easier to control.

The point I was making last year is that it was the IT-oriented folks who originally flopped in our world because they thought they would have the luxury of protocol and topology standards that they enjoy in their world.

And another thing … Although we've come a long way, it's not necessarily easy to match IP hardware with IP software.

Many home automation vendors still test various routers and PCs and operating systems, and advise, "Here's the ones that work best with our software; here's the ones to avoid; for the best result, use our PCs and interfaces."

An IP-oriented home automation system might not work for the simple reason that something in the software doesn't like a particular off-the-shelf router.

Wherefore Art Thou, Standards?

Ah, the eternal question. Let's get one thing out of the way.

Eloy writes:

In my opinion the world would be a much better place if we didn’t have all these different protocols (X10, Z-Wave, etc.) competing and fighting for supremacy and instead we’d have a single, open protocol on top of TCP/IP.

Perhaps in a decade or two when WiFi gets so ubiquitous that it can actually rival these low-rate protocols in price.

Yeah, it's lousy that we have ZigBee and Z-Wave and Insteon and LonWorks and countless other protocols are duking it out to win the hearts of home automators.

But TCP/IP cannot replace them. For starters, they are cheaper to implement--especially when it comes to wireless. Much, much, cheaper. Maybe 1/5 or 1/10 the cost? Just as important, they require less power, generate less heat, and have other characteristics that make them ideal for home-control applications.

It is the controller that communicates with these low-cost devices (sensors, thermostats, light switches, etc.) that should communicate via robust IP communications. You are correct, and we're getting closer.

But … to make it seamless, there still needs to be protocols for home control.

I believe it boggles every one in the industry that still no such thing exists, even after millions of dollars and decades of trying—CEBus, LonWorks (both EIA "standards"), UPnP, now Web Services for Devices.

LonWorks, the powerline control protocol that is very prevalent in the industry world, is making inroads into the home markets.

UPnP is still kickin' along. I believe the UPnP Home Automation & Security working committee recently updated its device control protocols (DCP) for controlling motorized blinds. The other major food groups are already represented in UPnP. But practically no one in the home automation knows about it.

Web Services for Devices (WSD) as a home-control protocol has enjoyed some popular appeal late, thanks to the resources of home-control vendor Exceptional Innovation, which is creating a library of definitions for different home-control device categories.

For now, like the others, EI has to "WSD-enable" third-party systems by running and hardware bridge and writing the same ol' stifling drivers.

More importantly, like UPnP, the home automation world today shrugs off WSD.

On IP-based A/V Distribution

Boy, back when I wrote the original editorial, I got an earful from Herman Cardenas, CEO of NetStreams which at the time made IP-enabled audio distribution systems. The company has now rolled out IP-enabled video (HD) as well.

I won't get into the merits of the quality of IP-based streaming. This has gone on long enough! There are limitations versus other distribution methods, but NetStreams and others have done a great job of addressing bandwidth limitations, latency, QoS, etc.

But still … not a whole lot of IP-enabled end points like TVs and loudspeakers, but they are certainly on the way.

What to Do Today

Run Cat 5+ everywhere you can -- multiple runs if possible. Include potential locations for keypads, in-wall speakers, wireless access points, telephones, intercoms, thermostats, surveillance cameras, etc.

Eloy is right that eventually these higher-rate communications systems (versus low-rate sensors and possibly light switches) all will some day communicate via TCP/IP. And, what's more, they will communicate with each other, and with IP-based interfaces, seamlessly and affordably.

Why H.A. manufacturers have failed in the past is that they assumed that already to be the case.

That's only one of the reasons, however. There are four more reasons listed as well!

Thanks for your indulgence. I welcome our comments.