Mesh finally comes to Bluetooth with major upgrade in range and rate

It’s been a long time coming, but the Bluetooth Special Interest Group (SIG) has finally released Bluetooth 5, the new version of the protocol that brings significant range and bandwidth improvements, as well as the new mesh networking features that will allow Bluetooth to challenge the likes of Zigbee, Z-Wave, and Thread in emerging IoT markets.

With around 92% consumer awareness, Bluetooth is likely only second to WiFi when it comes to brand visibility. It has a major advantage here over rival radio protocols, thanks to this consumer familiarity, and assuming the SIG’s messaging is on point, Bluetooth 5 and its mesh capabilities should enjoy considerable success in the smart home.

As long as smartphones continue to function as the main interface between a consumer and the digital world, Bluetooth will also benefit from its tight integration – allowing the smartphone to act as a local conduit between the user and their smart home devices. These phones don’t include Z-Wave, Zigbee or Thread radios, and the handset makers show no signs of adding them.

While Bluetooth doesn’t have a completely stellar reputation among consumers, still burdened by experiences with earlier versions, its performance enhancements, first announced in December, could resolve that contention. Headline improvements of 4x range or 2x bandwidth, to 200 meters or 2Mbps, are a big deal – although it’s important to note that these are an either/or choice, as you can’t have both at the same time.

A theoretical maximum range of 4.3km is not something that consumers are going to be encountering, but there are companies that are achieving ranges that make 200 meters sound conservative. Riot spoke to Cassia Networks a few weeks ago, on the back of a partnership with GE Digital, which specializes in using long range Bluetooth in industrial and commercial applications, using a few adaptations to the protocol to extend its range – but using standard end-device hardware.

It was good to hear that the brand name ‘Bluetooth Smart’ is now dead –its introduction was a little confusing, considering that Bluetooth Low Energy was the previous brand for the protocol, and far better suited to the IoT direction that the SIG was pursuing.

Ken Kolderup, the SIG’s VP of marketing, told Wireless Watch’s sister service, Rethink IoT, that the best way to learn about Bluetooth Mesh is to forget everything you know about Bluetooth – and a lot has changed under the hood. Explaining the protocol’s evolution, Kolderup noted that it had come a long way from its peer-to-peer audio origins. In 2010, the launch of Bluetooth Low Energy (BLE, in v4.0) saw the SIG focus on data transfer, rather than audio or voice.

BLE introduced broadcast (one-to-many) functionality to the protocol, which provided the foundation for beacons – which have gone on to be used in smart buildings, venues, retail, asset tracking, and as assistance tools for indoor navigation.

However, BLE required a new type of radio hardware, which is also the basis for v5.0 – meaning that Bluetooth Classic, the original radio, supporting BR/EDR (Basic Rate, Enhanced Data Rate), will still be required for Bluetooth devices looking to ensure complete backwards compatibility.

As Bluetooth 5 uses the same radio as BLE, the new protocol is technically OTA upgradeable – although many developers didn’t design their devices with such updates in mind, and so it is unlikely that most devices will support OTA updates from v4.x to v5. Any BLE radio should support Bluetooth 5.

Kolderup said that the introduction of mesh (many-to-many) rounded out the topology options for Bluetooth, but added that mesh is being targeted at new use cases, and isn’t trying to tread on existing Bluetooth applications. The envisioned large device networks, such as building automation and commercial lighting (big drivers for development, apparently) could not have used Bluetooth in the past, and so the SIG hopes that v5.0 will compete against the Zigbee and Thread 802.15.4-based protocols that have gained traction in those environments.

Interoperability is a core focus for the SIG and its new protocol, and is a core focus of the launch – with the three main tenets of the launch being “industrial-grade; proven global interoperability; and a trusted mature technology”. Kolderup added that the Bluetooth ecosystem should be seen as a value-add for mesh networking.

As for the more technical details, the mesh architecture defines Devices (products capable of being provisioned for the mesh network), and Nodes (a Device that has been provisioned for the mesh). The Nodes form a Network, and share the required encrypted network key. The Nodes communicate via Managed Floods, and use a routing system to ensure the messages are transported efficiently within the Network.

A basic flood just blasts a message to every device within range, and while very simple and pretty reliable, the basic Flood is rather inefficient in terms of battery life and power consumption, and can also be problematic through network congestion. Kolderup noted that people have a rightful concern over the term Flood.

The Bluetooth 5 solution to those concerns is in the Managed Flood, which provides time-to-live and message caching countdown functions, to remove messages from the Network after a set time or number of hops. Mains-powered devices can be set to act as Relays, free to pass messages between Nodes in the network.

Battery-powered devices can use the Low Power and Friends functions to prolong their uptime. A Low Power device can nominate a Friend device to store messages for it while it is asleep, so that when it is scheduled to wake up and listen to the network, the Friend device can pass on the stored messages. That sleep-time schedule can last up to four days, and there are options for sending both acknowledged and unacknowledged messages.

As for the size of these networks, a single one can have up to 4096 sub-networks, which are used to keep local traffic within that sub-net and not joining the wider network. An example of this might be zoned lighting, where different groups of bulbs communicate in local isolation from other groups, but are still listening for global commands. Notably, this architecture doesn’t require gateways to function – as the Nodes are capable of this internally.

The theoretical maximum size of a Network is 32,000 Nodes, supporting up to 127 hops per message, and the latency is being described as ‘speed of sound,’ rather than ‘speed of light’. Publish/Subscribe (PubSub) is also enabled, which provides multicast functionality for zones, such as lights in a hallway subscribing to their control switch, and the switch publishing commands for them.

However, as with all maximum claims, we will wait to see if any live deployments support such scale or speed. Kolderup says that there are 1,000+ node networks live currently, which are considerably larger than Zigbee and Z-Wave networks (with Z-Wave capping out at 232, and Zigbee networks often running into problems in the 200-400 range, based on our findings).

That 32,000-unit size does sound pretty excessive for applications like lighting or HVAC, but could be useful for facilities managers looking to create a single network for their range of applications, rather than relying on multiple individual networks for each different function. That choice will be dependent on the nature of the facility, and each project’s budgets, but it opens up campus-wide or large building networks for Bluetooth.