ZigBee and the evil radios

Ten years ago, Bluetooth, 802.11 and HomeRF were engaged in an acrimonious battle for supremacy over leadership as the short range radio standard.  HomeRF died, and in the following years Bluetooth and 802.11 found their areas of application and now coexist together, to the extent of joining forces in the new Bluetooth 3.0 specification.  Today a new and ferocious fight is taking place for the role of ultra low power radio champion.  This time, there is likely to be just one winner.


In the two main corners of the ring are ZigBee PRO and Bluetooth low energy (previously known as Wibree).  Alongside them, throwing lighter punches, are an array of lesser contenders, including Z-Wave, ANT, Wavenis, and Wireless M-Bus.  What is at stake is the prize of becoming the standard for connecting low power consumer products to the next generation of mobile phones and enabling smart energy devices within the home.


The industry has learnt a lot from the experience it has gained with Bluetooth and 802.11.  One of the first lessons was the need for interoperability between products from different vendors.  Interoperability isn’t something that comes from a specification alone – it needs a rigorous testing regime that ensures that different implementations work with each other and which check that every product entering the market is tested.  802.11 learnt that lesson.  It didn’t have that process until the Wi-Fi Alliance came along to complement the raw 802.11 specification and turn it into the usable experience that has made it successful.  Any new standard will need that from day one.


In practice, interoperability also needs a minimum of three different suppliers who can demonstrate interoperability.  Any less and manufacturers have little real choice of silicon supplier.  That reduces competition and increases risk, which stalls the market growth.  But to support three or more suppliers the market needs to be big.


Depending on its complexity, it costs between $5 million and $10 million to develop a complete radio chip with its associated embedded protocol stack.  If a vendor is going to be able to sell these at a price of $1 – $2, which is what the volume market will demand, then they need to be able to sell tens of millions to make money.  If we demand a minimum of three suppliers to ensure interoperability, that equates to a market of around 100 million devices.  Less than that and there will not be enough profit to support multiple vendors.  That means no interoperability, lower uptake and a declining future for a standard as a proprietary niche.  For the successful contender, the market can grow into shipments of billions, supporting multiple chip vendors and successive rounds of evolution.  It paints a very stark reality – that there will probably be just one such market to emerge of that size.  The winner will take all (or at the very least, the lion’s share).  That’s why the current battle is so heated.


There are two potential opportunities that offer the route to these volumes for ultra low powered radio chips.  It’s unclear which will be the first to emerge.  The first of these is smart energy.  To help limit the amount of power that consumers use in the home, governments are looking to replace utility meters with smart energy meters that have the ability to control domestic goods.  This allows them to be turned off at peak times, or to alert the user to an increased tariff, in the hope that this will modify their behaviour.  Governments around the world are starting to legislate for the introduction of smart meters.  In the UK, the Energy Retailers Association has been more advanced than most in drafting a plan to assess the different wireless technologies and will shortly start field trials.


The second market is for connected health and fitness devices.  Changing demographics are opening up a market for connected health, whether that is for general fitness, long term chronic disease management or assisted living.  Once again, a low power wireless link is seen a key to deployment, connecting devices to mobile phones and home gateways.  The Continua Health Alliance is actively developing standards to help ensure compatibility of data from such devices and is expected to choose a candidate low power radio in the near future.


The standards needed to meet these demands are new.  Unlike existing low power standards such as Bluetooth and the earlier versions of ZigBee, they need to run off coin cell batteries for months or years.  As yet the main contenders are largely untried – they build on existing standards, but the specifications targeting these applications are still at the final stages of development.


In this atmosphere, the political temperature is rising.  There are probably only two serious contenders which have the potential to reach the volumes necessary for success – ZigBee PRO and Bluetooth low energy.  The reason for this is that they are the only two which can claim a route to get to the hundred million mark.  For Bluetooth low energy it should be easy.  One of the clever aspects of the standard is that the Bluetooth low energy technology is incorporated into the next generation of normal Bluetooth chips, so an increasing number of mobile phones will incorporate it for free.  That gives a critical mass of hundreds of millions of handsets that can act as gateways or displays for a new generation of products.  So whether or not the smart energy or Continua Alliance supports it, Bluetooth low energy will gain market traction.


For ZigBee PRO, the base 802.15.4 radio chip is not used just for ZigBee variants, but for a range of other standards, including 6lowPAN, ISA100 and wireless HART.  At the Wireless Congress in Munich last year, a number of chip suppliers estimated that all flavours of ZigBee may account for less than 25% of 802.15.4 usage, so even if ZigBee PRO grows slowly, it can ride on the volume created by other applications.

 802.15.4 Chipset Market


However, that’s not necessarily an advantage, as it only covers the lowest layers of radio silicon and baseband.  Vendors still need to write the protocol stacks for ZigBee PRO and if a semiconductor manufacturer is selling their 802.15.4 chips into multiple markets, they’re less likely to do so, leaving the work to external partners.  In addition, the chips are less likely to be optimised for the requirements of ZigBee PRO, as they will need to cater for a broader range of applications.  That may be bad news for price as well as interoperability, in contrast to the integrated approach of Bluetooth chipset vendors who do the complete job in a single chip.


Because of its integration into handsets, Bluetooth low energy is assured of success in reaching critical mass.  In contrast, ZigBee PRO needs to win the approval of key industries if it is to survive.  At a recent conference a spokesman for one of the major chip companies behind the ZigBee PRO healthcare profile let slip the fact that if it failed to win Continua approval, his company would probably kill their ZigBee PRO healthcare development effort.  Other proponents within the ZigBee PRO camp are openly talking of the competition with “evil radios” as they ramp up the level of rhetoric (and give this piece its title).  That smacks of an increasing level of desperation, but then again, ZigBee needs this approval far more than Bluetooth.


What is clear is that the smart energy and healthcare bodies need to make sure that they base their decisions on hard facts.  The adoption of a radio standard will have a long term bearing on the performance and availability of desperately needed products.  That choice needs to be firmly based on technical merit, not bestowed as an act of charity.


The next few months should be interesting.  The support of major handset suppliers will ensure that Bluetooth low energy has a future.  If that argument of volume and associated development resource wins the day and Bluetooth low energy is selected for healthcare and smart energy, then 2009 could be a painful year for many of the other aspiring low power radio standards.




I am afraid I have had to disable comments on this post, as someone has been bombarding it with spam.  It’s sad to think that someone in the technical community is unable to cope with honest debate.  I am sure that their action is neither appreciated nor endorsed by the majority of those working on ZigBee.