Z-Wave Goodbye!

Silicon Labs is buying Sigma Designs, practically the only vendor offering chips and stacks for Z-Wave, the established home automation standard. It is an interesting move and watching what happens next will be even more interesting. Contrary to what the press release says and what other people are saying, I think it is the end of the game for Z-Wave. Z-Wave has been the longest standing and most complete home automation standard. Mainly due to the fact it has been tightly held by a single company (Sigma Designs) who has been defacto the only supplier of silicon and stacks for Z-Wave devices. Achieving interoperability among a family of devices coming from a single vendor is easy. And that is why Z-Wave has been so far the only option for interoperable smart homes.

But Z-Wave as a technology has been ageing quickly. It does not scale to cover a home full of smart devices. It has very weak or no security. Majority of Z-Wave devices on the market today speak clear text to each other, secured with only an 8-bit CRC check. There are advanced security modes proposed by Z-Wave but almost nobody implements them today. And even they have shortcomings: it is possible to use secure Z-Wave only for unicast traffic: the receiver sends a nonce to the sender for the latter to use it to secure a message. A mode not applicable to multicast topology, meaning there will always be a popcorn effect, even when controlling a handful of lights. The frequency bands are fragmented and affected by sub-GHz LTE rollouts. There is a source-based routing scheme that takes very long time to heal a broken route. And many other problems, collectively almost impossible to fix. So let's face it: Z-Wave is dead.

There is one jewel though, worth considering: Z-Wave has a very complete and nicely engineered application layer. It covers hundreds of device types, diving deep into their specifics. Maybe not as deed as Bluetooth mesh does for lighting and sensing, but still fairly deep. This application layer could be decoupled from the underlying Z-Wave radio and run very effectively over other, more modern transports. Such as Thread or Bluetooth mesh. While for Thread this would probably mean more fragmentation and more question marks (why would we need yet another application layer?), SiLabs could take advantage of it (and engineers behind the design) and run it over Bluetooth mesh. Which would be a very interesting development. Provided a Z-Wave compatible stack is offered on top of Bluetooth mesh, this would enable many manufacturers of smart home Z-Wave devices to upgrade to Bluetooth technology, solving instantly the scalability, security and performance issues plaguing Z-Wave today. And would put SiLabs in a position to offer pin-compatible and function-compatible Bluetooth chips to Z-Wave vendors.

Looking forward to the next move fro SiLabs!