True Interoperability

The news of the first batches of commercially available Bluetooth mesh devices working seamlessly together have circled the world recently. This is a powerful demo indeed. Two independent products from two competing companies who never met before, taken off the shelf, work seamlessly together.

It seems simple, while it is not. Actually, this is the first time ever, it happened.

The details can be traced down to the two Bluetooth listings: the D038467 by LEDVANCE GmbH (who make the Sylvania light bulbs based on the Cypress Semiconductor module and stack) and the D039781 By Nordic Semiconductor (who offer a Bluetooth Mesh Profile subsystem loaded on the NRF5 development kit). These are totally independent implementations of the Mesh Specifications. And the products "just" work together.

Some may say we had this before, with Z-Wave. This is true, but only to some extent. Z-Wave really succeeded as a standard facilitating interoperable products, but with two caveats:

1. All Z-Wave products have been using the same silicon and the same software stack: both from Sigma Designs. There were no alternate implementations. So we may say that Z-Wave was effectively interoperable with itself.
2. To interoperate, Z-Wave products had to come from the same geography. Z-Wave uses different frequency bands for different geographies, so for the products to work together, they must come from the same region. A US-based Z-Wave light bulb will not respond to commands from an EU-based Z-Wave switch, as they cannot hear each other.

There are a number of things that make Bluetooth mesh unique and facilitate this true interoperability. Among them are:

1. The globally unified frequency band. Unlike other wireless technologies, Bluetooth has no borders. It works in any country and there are no variants of products for this or that market. For any product there is a single SKU that is global.
2. The strictly defined and tested application layer, the Bluetooth Models. I cried for that years ago. Skip Ashton was echoing me in 2016. The whole lighting industry was raising this need on many occasions. Bluetooth mesh delivered on this front in 2017 (for lighting, sensing, time and variety of generic applications) and keeps broadening the scope (e.g., via the smart-home focused effort).
3. The strict definition of - what I refer to as - Layer 0 of the OSI stack - the Onboarding layer. In IP/Ethernet the layer 0 is the RJ-45 Jack format and knowing how to plug a cable into a switch. In wireless it is about how to bring devices into a network, which involves discovering what devices are available and what are their capabilities. Bluetooth Mesh calls this 'Provisioning' and it is a separate stack, defined by the Section 5 of the Mesh Profile specification. It is the most omitted part of most standards that try to claim being interoperable. Yet the part that no products can be set up to interoperate without.

True interoperability is not a goal nor a milestone. It is the cornerstone: the mindset and the process. A mindset that prioritizes all actions necessary to make products interoperable and a process that ensures this is really happening. In Bluetooth, interoperability testing starts at specification development. All specifications that have been adopted and published had gone through thorough interoperability testing before adoption. And then the products, before release, go through cross-vendor massive testing at regular UPF (UnPlug Fest) events.