22 Comments

Summary:

The wide array of wireless radio technologies used to get devices online may soon shrink as major players in the chip world start choosing the standards they will support for the internet of things.

WiFi signal
photo: Shutterstock / Jiri Hera

There is nothing that hardware nerds love than a good old-fashioned standards battle. LTE versus WiMAX, VHS versus Betamax, Ethernet versus InfiniBand … the list goes on. The internet of things is another battleground with different factions fighting over protocols for sending wireless signals, sending data between points, security and a variety of other standards. But when it comes to the wireless technology of choice for connecting consumer gadgets, Broadcom has chosen its winners: Bluetooth and Wi-Fi.

Brian Bedrosian, Broadcom’s Senior Director, Embedded Wireless and Jeff Baer, Broadcom’s Business Development Director, Embedded Wireless — both in Broadcom’s Wireless Connectivity Combo Group — explained their thinking to me on a call related to the launch of a new all-in-one Wi-Fi module that contains a Wi-Fi radio and a microcontroller that will handle all the on-boarding of the device and communication to the network.

From Broadcom’s perspective, other protocols are either closed or the standard is so open to interpretation (Zigbee) that it might as well be closed. For example, there’s no guarantee that Zigbee devices will work with other Zigbee devices, and for Z-Wave, the chips are more expensive. Plus, neither Zigbee or Z-Wave are regulars in the smartphone radio stack. I’ve discussed this with the CEO of Securifi, the maker of the Almond + router, on one of my podcasts, if you want to learn more.

Thus, Broadcom has made its bets on Wi-Fi and Bluetooth (and Bluetooth Low Energy). That means we can expect more modules from Broadcom that make it easier to build connected devices, something the Bluetooth Special Interest Group is also pushing hard. For example, the latest chip, the BCM4390, is designed with a Wi-Fi radio and a communications processor so the whole package can go into existing devices like washing machines or dishwashers without requiring a separate microcontrollers to handle the additional burden of networking. That means devices can be retro-fitted for connectivity without swapping out other parts.

Broadcom was coy on the subject, but I imagine a Bluetooth and a Bluetooth Low Energy module that can handle the networking on-boarding isn’t too far behind. It did launch a Bluetooth system on a chip today as well, that offers much lower power consumption so people can go a year without needing to replace the battery in connected device like a pedometer or door lock. And by the way, this isn’t all just to boost the consumer experience. As more and more businesses are realizing, adding connectivity can help them, even if that connectivity isn’t exposed to consumers through fancy apps or whiz-bang refrigerators that tell you when the milk is expired.

Baer explained that businesses want connectivity so they can track data on how products are used, or update features over the air. So even if the microwave doesn’t need an app, it may need Wi-Fi. And Broadcom really wants to sell that Wi-Fi chip.

  1. Maurizio Melchiorre Wednesday, May 29, 2013

    but this new chip wifi will support a Wireless mesh network?

    1. Broadcom says, ” The new BCM4390 chip does not currently support mesh with WLAN. Because the home Wi-Fi network offers whole home coverage and extra nodes/set-up can introduce additional costs and complexity, a simple range extender is a more optimal solution for consumers requiring additional range for hard-to-reach spots in the house or basement”

  2. Wifi and bluetooth networks both get worse each time you add another device. Zigbee networks get better. Zigbee needs to get on the ball and get more product out on the open Market if it wants to win overall though.

  3. Good article. Zigbee was always a bad idea (there should not be any one single company or entity to dictate a protocol). I’m glad to see this move by Broadcom adopting open standards, open protocols (long live WiFi and Bluetooth 4.0+).

    1. I think you meant Z-Wave there. Z-Wave is owned and “operated” by a single company Sigma Networks (purchased Zensys, original company behind Z-Wave). ZigBee on the other hand is a truly open IEEE standard based on the tried-and-true 802.15.4 radio standard in the 2.4 GHz band. There are number of large viable companies that make ZigBee chipsets including Silicon Labs (formerly Ember), Texas Instruments, Freescale, Greenpeak, and others. ZigBee also has a significant battery life and range performance over both Wi-Fi and Bluetooth – some of the better ZigBee sensors have 7+ year battery life from a single CR2 battery. Wi-Fi and Bluetooth are great for many purposes, but Wi-Fi is terrible at battery-powered/low-power devices and Bluetooth has serious range issues (ZigBee devices can easily be 2x Wi-Fi and 6x Bluetooth in terms of range). This reads like a marketing press release from Broadcom to play to their existing strengths.

      1. What a about the claim that zigbee devices are not compatible with each other? Is it true?

        1. just a visitor joe Thursday, May 30, 2013

          well, did you see any proof of that statement in the article? Monique calls it right I’d say: marketing press.

          1. Stacey Higginbotham just a visitor Friday, May 31, 2013

            Actually, the Nest is a good example of the incompatibility of Zigbee in that it uses Zigbee, but it doesn’t pair to other Zigbee devices in the home. Bluetooth had similar issues in the early days and was famously declared “dead” only to solve those problems and clearly thrive.

            1. I agree that Zigbee devices from different suppliers have compatibility problems, no doubt about that. Two reasons, IMHO, the spec is implemented by many companies, and secondly much of the Zigbee functionality in the early days was implemented using extensions or customization.

              But I am confused that the Nest does not pair with Zigbee devices – that statement will hold true if and when Nest opens up to the fact (officially) that they have Zigbee and now they are going to support interfacing to Zigbee chips. They could or probably have not enabled Zigbee function to the consumer.

              Nest 1 had the TI 2530 Zigbee with TI CC259x RF Front end (a combination similar to found in Philips Hue bridge), the Nest 2 migrated to the Ember/Silicon Labs chip with a Skyworks front end. Ember happens to be an early pioneer in Zigbee – I don’t see any reason why Nest cannot work with Zigbee devices.

      2. And that’s part of it’s problem, as each one of those companies make proprietary Zigbee stuff. Additionally I find it hard to believe they have a 7 year battery life, when the battery itself won’t last more than ~2.

      3. Jonathan Adams Monique Friday, May 31, 2013

        ZigBee is not an IEEE Standard. The 802.15.4 radio ZigBee uses for 2.4GHz is a an IEEE standard. The ZigBee protocols (yes plural and often incompatible) are owned and maintained by the ZigBee alliance.

      4. Zigbee is standard on paper only as there is no interoperability certification. This was also a problem with WiFi and Bluetooth in the early days, so Zigbee can still address this problem, but honestly after 15 years, you would think they would have stepped up. They had their chance as far as I am concerned.

        After almost 2 decades of fighting with everything from x10 to Z-Wave to Zigbee to even PLC, I am happy to see an easy and standard way of networking home automation that is truly accessible and hoping that we are finally close to achieving a dream I have had for a very long time.

  4. The profile for devices and sensors is diverse which means this game will be played for long time to come

    WiFi is great as long as the device has direct access to the power line. Not so much for sensors. Interference is an issue, that data won’t be moving while the microwave is on.

    Mesh networks are resilient but problem is that to build a mesh you need many devices. A bit of a chicken and egg. You can’t try building a mesh with a single device.

    Finally there’s the whole thing with batteries. Changing batteries to door locks every six months? That will get old pretty fast. Low energy for BT, DECT etc will help. But a whole another area is energy harvesting. And the power constraints there are likely to establish yet another family of protocols.

    So the battle of radio waves will continue for quite some time. The future of RF deployment may end up being heterogeneous. And that can cause headaches.

  5. I guess this is more than true. There is no such thing like “THE” Zigbee but various dialects of Zigbee. Thats kinf of frustrating. Some 2 cents to Moniques comments.
    – Zigbee is provided by various chip vendors but not even they aork well together
    – Z-Wave is provided by two vendors (not one) – no idea how well they work together
    – Z-Wave is also an open standard, ITU-R G.9959 . You can get the full spec from the ITU in Switzerland.
    Clearly both zigbee and zwave have way better battery life time tan WLAN but this statement is not true in comparison to bluetooth

    1. Dirk – Do you know who the second supplier is? Sigma Designs who bought Zensys is the only supplier I am aware of. It would be nice to know if indeed they have second supplier – and if the second supplier has designed their own Z-Wave chips or simply licensing and re-selling the Sigma/Zensys IP.

      1. Stacey Higginbotham Ashu Joshi Monday, June 3, 2013

        Ashu, I believe Sigma, since it purchased Zensys, is the only Z-wave supplier on the market. I’ve never encountered another supplier, although perhaps there is a grey market company operating out of China or someplace?

  6. Broadcom is not the only player in the market of providing a WiFi chip with Communications Processor for appliances. Marvell for example (http://www.marvell.com/smart-energy/assets/Marvell-Smart-Energy-Platform-Brief.pdf) has chips in this segment and so do companies like TI. WiFi is fantastic for powered end nodes, the technology today simply is not suited for battery powered sensors. I have been, for example, trying to play with the Twine – and being WiFi – running it in batteries for connecting things is messy – the 2 AA batteries keep dying on me.

  7. BT has low battery/power consumption but not enough RF power to cover a home. It’s a PAN.

    Wi-Fi has decent home coverage (although not great when devices are screwed into wales, doors, corner etc.) but indeed have a high battery/power consumption. It’s a power hungry HAN.

    The IoT has surely a number of application areas where you only need coverage OR low battery/power consumption. This is where Broadcom’s strategy may work.

    In the home, office, SoHo, SMB, you need full home coverage AND, since ~50% of the devices are battery driven, AT THE SAME TIME you need a low battery/power consumption. This is where BT & Wi-Fi fall short.

    BT and Wi-Fi (and many other star based networks) are fundamentally not designed for this combination of full home coverage combined with low battery/power consumption. I am sure that Broadcom realises this but if you only have a hammer, all problems look like nails.

    Further BT and Wi-Fi do not have profiles for IoT. It took indeed a long time to get resolved in the early days of BT and Wi-Fi. Doing this in sensors, locks, light switches etc. will certainly not makes things any easier and will take a lot of time to resolve. This interoperability is also where Zig-Bee is performance is weak and Z-Wave is particularly strong. Companies that are operating in this IoT market have discovered this hence the very broad support of Z-Wave in the IoT industry.

    1. Stacey Higginbotham Raoul Monday, June 3, 2013

      Raoul, the Bluetooth SIG has been really pushing the range on BT so it covers an entire home. The range exceeds 100 meters for standard BT according to the SIG and it is working to push that further. I was shocked to discover that, actually.

      1. Stacey, the theoretical range of Bluetooth may exceed 100m, but in practice there are many factors such as antenna design (which, sadly, is often an afterthought) and the objects in the environment between the transmitter and receiver which significantly reduce that value. “Tens of metres” is a the best description we’ve seen, and it’s the one we use ourselves.

        Raoul, totally agree with your analysis. That’s why at reelyActive we’re making inexpensive sensor infrastructure that supports multiple standards and that you can easily daisy-chain to extend the coverage for a SoHo, SMB and spaces much, much larger. And, as a bonus, you get point-of-interest location, based on the sensor which decodes the strongest signal. Bluetooth Low Energy is arguably the most interesting (existing) standard once you overcome the range limitations.

  8. Tolga Latif Friday, June 28, 2013

    Would you please clarify a point. You wrote:

    “Thus, Broadcom has made its bets on Wi-Fi and Bluetooth (and Bluetooth Low Energy). That means we can expect more modules from Broadcom that make it easier to build connected devices, something the Bluetooth Special Interest Group is also pushing hard. For example, the latest chip, the BCM4390, is designed with a Wi-Fi radio and a communications processor…”

    What are you referring to as a module in this context? In my understanding, a module would be the SOC or RFIC and Micro with the crystal filter and matching circuit. You wrote that we can except more modules, but then gave the example of this case as a SOC, the BCM4390. Is Broadcomm moving into modules or can we expect more SOCs (RFIC + Micros) from them?

    Your clarification would be appreciated

    1. Stacey Higginbotham Tolga Latif Friday, July 12, 2013

      I was referring to the SoC. Sorry for the confusion.

Comments have been disabled for this post