Will Google kill ZigBee?
- in Wireless
Or will it main Bluetooth? Or Wi-Fi? Or maybe Z-Wave? Or any of the multitude of other short range wireless standards. It’s a question that was raised last week when Google did a keynote presentation on Android@Home at their I/O Conference where they announced a wireless light bulb which could be turned on and off from a mobile phone. The technical details are very sketchy – much of it coming from Lighting Sciences Group, who did the accompanying demonstration. It’s unclear whether it’s a new radio, a new protocol, a new standard or even what frequency it’s running at. But you don’t expect the absence of little details like that to stop speculation.
The greatest level of speculation has come from the smart energy industry, who are suggesting that ZigBee could be the main casualty. Jesse Best at Smart Grid News asks whether this will take away ZigBee’s momentum. And there’s an interesting range of comments about that on his site about that, which are worth reading. Throughout the industry, Google’s announcement is making people question whether they’ve made the right choice?
I’m not sure that anything Google does will displace ZigBee from its place in smart meters. That’s actually quite a closed market, as most utilities don’t really want to share that data with consumer devices. Where it is a threat is in home automation. Home Automation is still a very nascent market and Bluetooth, Wi-Fi and ZigBee are all pitching to own it. The reason I think they are at risk is because of what Google can bring, which is an API (Application Programming Interface). Google has succeeded in areas like mapping because it makes it easy for developers to access and mash up data. In contrast, wireless standards shy away from making their stacks easy to use, particularly for embedded designs. If Google can make it easy, thousands of garage and backroom developers will take it and innovate with it, and the existing standards may all find themselves left behind.
It’s worth pointing out that Google has some major challenges if it is going to achieve any level of success with a wireless stanard. Robust wireless standards are not easy to write. Bluetooth, Wi-Fi and ZigBee are all ten years old – it’s taken that long to get them right. Wireless is difficult – it introduces issues of security, topology and connectivity that aren’t there is a cable. For more on the problems of writing a wireless standard, read the book.
That made it odd to hear Eric Holland of Lighting Science Group expostulating about their shortcomings, citing interference, latency and difficulty in deployment. He may know something the rest of us don’t, but it sounds like a glib marketing story. According to reports the radio will “run in the 800 – 915MHz band”. That is actually a lot more limited than it sounds, allowing a limited number of devices. Plus it’s not a global band, so products designed for the US will be illegal in Europe and vice versa. There are already other standards working in that range, so they may be building on one of those, but most of them are pretty basic and need a lot of work if they’re going to grow to support billions of devices. Which takes a lot of time. The low frequency has the advantage of having better range than 2.4GHz, which is good for home automation. But it has a significant disadvantage which is that it requires yet another antenna for phone and tablet manufacturers to add to their devices, which may make it difficult to persuade them to adopt it.
Then there’s the issue of regulations. However clever you are, you can’t just design a new radio and sell it. You need to get it approved in each country you sell it and meet some very strict legal requirements on how your radio operates. There are ways around that; some manufacturers make certified modules, which can be used in products without the need for further certification, but it’s a process that is highly regulated and poorly understood by most home automation developers.
Plus there are IP issues. Holland claims that this is a mesh network. If it is, it’s probably infringing someone’s intellectual property – quite possibly that of the ZigBee Alliance members. If it’s using security, then it’s probably infringing IP from within Bluetooth and Wi-Fi. And if it’s not using security you don’t want to use it, or else your neighbour’s kids will start controlling your lights. Moreover, the people who own much of this IP, such as Nokia, Ericsson and Qualcomm, have bigger and nastier lawyers than Google.
My guess is that this is an early demo, with a partner who has given a rather naïve press interview. Nevertheless, the industry should not underestimate it. As I said earlier, Google understands the value of APIs. If it chooses a competent wireless standard and provides the right APIs it could quite easily capture the home automation market, killing it off for anyone else. And possibly discard Lighting Science Group along the way as an expendable means to an end. Alternatively Google could decide to sit an API on top of a number of different standards, and let the market decide on which one to use. Although the user experience from that approach is not wonderful.
The alternative is for the competing wireless standards to make their standards easier to use, but I’m not confident about that. I’ve sat in standards meetings when working group members have argued that it’s not there job to write APIs, because that’s not what a wireless standard is about. It’s as if they’re afraid that someone might actually use their standard if they make it easy to use.
So it will be interesting to see what happens? In a soundbite fixated world this announcement will increase the level of chatter which can only be negative for Bluetooth, Wi-Fi and ZigBee. With the smallest market base of the triumvirate, ZigBee probably has most to lose. If it gets its head down and concentrates on getting products out it should be fine (which should be a wake up call to those in the current SEP2.0 debate). But it’s a diversion that ZigBee could do without. On the other hand, if the industry delays its smart meter roll-out, all bets could be off.
Bluetooth may not fare any better. It’s finally getting to the point where Bluetooth low energy may see the light of day after a tortuous journey. I’d suggest that the Bluetooth SIG should be offering that standard to Google on bended knee, as Bluetooth low energy desperately needs some momentum to get it going. So Google may be the best bet for Bluetooth low energy, if the Bluetooth potentates can bring themselves to descend from their Ivory Tower in Seattle. There is alreadysome confusion about what the underlying wireless offering is in Android@Home. Alongside Holland’s assertion that it will be a mesh in the 800 – 915MHz range, other commentators have reported that Google’s plans do include Bluetooth, so maybe that journey from Seattle’s Olympian peaks is already taking place.
Meanwhile, Wi-Fi may find its aspirations outside wireless internet and video streaming dispatched to an early grave. Which is worrying for those companies trying to carve out a niche for low power Wi-Fi applications.
There’s an interesting new partnership opportunity which could threaten them all, which is for Google to pair up with low power DECT. That’s having something of a resurgence and offers a radio and protocol that would make a rather good match. But it’s probably a little too bizarre a partnership to contemplate.
So, its time to put down you bets on Home Automation. It will probably be a long and dirty path. And unless you plan to replace any purchases within the next three years, it might be a good idea to put off buying anything until then. Just remember how much good getting up and walking to the light switch is doing you. If we’re lucky, or maybe it should be unlucky, we may be the last generation to do that, so that it’s a journey our children will never take…
My guess is the vast majority don’t. But attitudes change. Ten years ago most of us never have contemplated mobile banking or giving up on real books. A lot will depend on how it’s sold and how easy it’s made. If it sells to the fear of “did I leave the iron / oven / lights on?” then it could take off sooner than many anticipate.
I do think there’s a market for more intelligent climate control which could learn your behaviour, with the phone as a personal controller to start warming or cooling the house when you’re on your way home. That needs designers to think about the balance of cost savings versus usability, designing for the mass market rather the the technorati. And I don’t see too much evidence of that yet.
There are some un asked questions… Do people really want to control their house with a phone?
Even with security we have seen hackers, virus and malware run amock on the internet.. so do we want to make it simple for them to move into our homes as well. I can just imagine the havoc caused when a virus takes over your home automation 🙂
I didn’t think I was arguing against other frequencies. I’ve always felt that 2.4GHz is not the best spectrum for smart meterings, as you see back in a blog from 2009. However, a lot of governments and other standards groups want to use what they believe is an established standard and ZigBee has done a very good job of convincing them that it fits the bill.
My comment was that the phrase 800-900MHz band is rather naive. As you point out, there is a limited amount of spectrum available at 868MHz in Europe, and a larger chunk of specturm around 915MHz in the U.S. In Europe today, many energy monitors operate at the lower band of 433MHz, which gives even better range, and still has a perfectly acceptable data rate for smart metering.
The point I’m trying to make is that there is a lot of competition for the home automation market, and ultimately only one standard is likely to dominate. It will probably not be determined on the merit of any protocol or frequency. The determinimg factors are likely to come from phone and tablet manufacturers. They will probably choose 2.4GHz because they don’t want to add another antenna and radio. And the protocol will be determined by whoever gets there first with a decent API for Android.
This article has a lot of misinformation — particularly about the 800-900MHZ. Any 800-900 MHZ radio can be used in Europe using the 868 MHZ frequency. The 900 MHZ mesh radio can have up to 65,000 individual nodes, and can transmit up to 256K data packets. There are plenty of 800-900 MHZ radio in use now within the utility –most legacy systems (AMR) are 900 MHZ. Do I need to continue? Please check your info before making statements that give the impression that ZigBee or any other RF protocol is better. No protocol is better –they just are used differently based on needs.
I repeat my statement that we are probably ten years too late for a licence free radio link protocol. Or at least one that includes any realistic mechanisms for advanced coding (whether that’s for speed or range), robustness. topology management and security, as the patents have already been granted. In ten to fifteen years time, when they expire, that may be different, but for now, the safest way to implement wireless is probably to sign up to one of the current standards.
> Will Google kill ZigBee?
Sorry it will not, because ZigBee is already death.
In week 3 of year 2011 I showed my presentation “Smart Car and Smart Home by Smart Phone”. It is showing all wireless links that are covered by a Smart Phone of today:
NFC, Bluetooth, ANT+, WIFI, GSM, GPRS, 3G and last but not least 6LoWPAN. 6LoWPAN is just a wireless UDP/IP communication. ZigBee was not included in the presentation, because there is no Smart Phone that is supporting ZigBee. BTW, the 6LoWPAN module I am showing is supporting 868 / 915 MHz.
Anyhow, the drawback of ZigBee and 6LoWPAN is need of powerful MCUs with a lot of memory.
In the meantime the big German energy company RWE has started with advertising the smart home.
They are using the 868 MHz band as well. In December 2010 the popular German computer magazine Chip published an article about the smart home. If you are following the link to Kooaba http://my.kooaba.com/periodical_pages/180850 then you get the text and the drawings. Kooaba is a new service to share articles legal. The test at the link is in German language, but the pictures will tell more than thousand words.
In Germany we use Wired M-Bus and Wireless MBus for to read the meters.
The Wireless M-Bus is on 868 MHz as well.
Rest in peace ZigBee. Welcome Google. Welcome openness. Welcome to licence free operating systems or radio link protocols.
I’m afraid that’s a common misconception. If you want to certify a Wi-Fi product, then you need to be a member of the Wi-Fi Alliance. Even at the lowest adopter level, you still need to sign an agreement which covers IP rights.
Unless you sign the agreement, you are not entitle to market your product as a Wi-Fi certified device. There are some procedures to allow rebranding of certified products by another company, but these are limited.
The same process also applies to Bluetooth.
Essentially, if you want to manufacture and sell products conforming to the Bluetooth, Wi-Fi or ZigBee standards, you need to sign an IP agreement. However, you should not ignore the benefits that this provides. There are a lot of patents in this space, and memebrship of these groups gives you access to a patent pool which significantly diminishes your risk of infringement. It is important to remember that open source does not equate with lack of IP infringement. That is a common mistake made by open source companies entering the wireless world.
As opposed to Zigbee a commercial Wifi developer doesn’t have to be part of the WFA (Wifi Alliance), a company can have a certified Wifi solution within their product without becoming a WFA member
There is a reason why ZigBee takes this approach, which is broadly similar to that of Bluetooth and Wi-Fi, which is that radio protocols contain a multitude of patents, many of them fundamental ones. When wireless standards are developed, patent holder agree to allow other members of the standards group to use them as part of that specification, which allows manufacturers to use that standard without risk of being sued.
You can’t wind the clock back and take those patents away. And it’s unlikely that any open source radio standard will emerge because most of the underlying technology that is needed for a secure, robust network would infringe some of those existing patents.
So the best route would be for Google to work to provide APIs on top of one of the existing standards. Open source developers would not design radios from scratch – it costs around $3 million to design a radio chip. Instead they’re more likely to use a pre-qualified module. And for some standards you can use those without being a member, as the license is part of the module price.
The alternative for those wedded to GNU is probably to stick with pieces of string and tin cans. They patents for those have expired.
Going back to the serious point, I think it would be a really good idea for Google to put an API on top of either Bluetooth low energy or low power DECT. Both should work with the pre-qualified module approach, so developers can just take them and use them. And both are simple to use.
Google is reluctant of using Zigbee due to software licensing issues, Google is using Apache license for every Android related technology and software and I presume they’ll do the same for there 802.15.4 based Zigbee substitute, See the following about Zigbee conflict with free software licenses (from Wikipedia):
The click through license on the ZigBee specification requires a commercial developer to join the ZigBee Alliance. “No part of this specification may be used in development of a product for sale without becoming a member of ZigBee Alliance.” This causes problems for open-source developers because the annual fee conflicts with the GNU General Public License. From the GPL v2, “b) You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.” Since the GPL makes no distinction between commercial and non-commercial use it is impossible to implement a GPL licensed ZigBee stack or combine a ZigBee implementation with GPL licensed code. The requirement for the developer to join the ZigBee Alliance similarly conflicts with most other Free software licenses.