Bluetooth: A technology in transition

 This is yet another episode of the “Wireless protocols showdown”, and this time we are reviewing Bluetooth, or to be more precise, Bluetooth Smart.

The history of the original Bluetooth standard started back in 1994, which makes such dinosaurs as ZigBee or Z-Wave look pretty young. One might wonder why such an ancient technology is even being considered by us here. After all, when reviewing Z-Wave, we concluded that this 12-year-old solution has failed to keep up with the rapid evolution of the IoT – so how on earth could Bluetooth be a viable choice if it was first developed before the term “Internet of Things” was even coined? A short answer is that Bluetooth of today is something completely different than Bluetooth of the past.

Bluetooth OSI model

The original Bluetooth, known as Bluetooth Classic, was designed as a short-range, cable-replacement technology for point-to-point communications. Initially, the main goal was to synchronize data between mobile phones, but the standard quickly became a default technology for wireless data exchange between personal computing equipment (mobile phones, PCs, PDAs) and various peripherals (headsets, cordless keyboards and mice, printers and such). Devices could form a tiny personal area network (PAN) called piconet, where a single central device would coordinate the activity of up to 7 active peripherals. Bluetooth Classic certainly did the job it was intended for, although we’ve all probably had some mixed experiences with it. Especially in its early days, the user experience was far from ideal, the process of pairing devices would sometimes end up being a bit frustrating, and the transfer of data would drain the battery very quickly. But that’s past.

Fast forward to 2010, the Bluetooth Core Specification version 4.0 is released, introducing Bluetooth Low Energy (BLE), more commonly known as Bluetooth Smart. This is where the story of Bluetooth in the Internet of Things really begins. Bluetooth Smart was designed specifically to address the needs of a new generation of smart devices, many of which are battery-powered and thus require fast connection times and efficient power management to reduce unnecessary energy consumption. This extended Bluetooth’s usefulness to a whole new range of products, ultimately making it a go-to technology for all kinds of wearables. And while Bluetooth Classic still remains part of the Bluetooth Core Specification, it doesn’t have much use in the IoT universe – so let’s forget about it for a while. Whenever we mention the name of Bluetooth from now on, we’ll mean Bluetooth Smart.

Interestingly, even though 4 long years have passed since Bluetooth Smart made its debut, there is still a lot of confusion in the market about what it is capable of, and what applications it fits. And rarely is it even considered an option for smart home or smart office environments. There are reasons for that, and we’ll try to explain both what these reasons are, and where the Bluetooth technology stands at the moment. Traditionally, the simplified OSI model is what we’ll begin with.

Bluetooth OSI layer

Pretty much like Z-Wave, Bluetooth covers all of the layers of the primary reference model for network communications, from the physical layer up to the application layer. So the Bluetooth Special Interest Group (SIG), the body which oversees the development and licensing of Bluetooth, has the rare privilege of being able to introduce any modifications to the standard directly and independently. This allows the SIG to respond to market developments and customer needs in a timely manner, making Bluetooth way more agile than other leading wireless communication technologies. We’ve already explained why e.g. the ZigBee Alliance doesn’t have this comfort of independence, so take a look at our blog post on ZigBee if you’re curious about that. We’ll get back to the SIG later on, now let’s focus on the protocol itself.

Out of all low-power, low-bandwidth communication standards, Bluetooth has the best radio. Period. This might seem like a bold statement so we’re certainly open to discussion, but the facts really do speak for themselves. Not only is it the fastest protocol, capable of transferring data with a rate of 1 Mbit/s, but it is also the most energy efficient solution out there. Why the throughput matters when smart devices usually broadcast very simple commands or only tiny bits of information? Because a higher data rate means lower duty cycle, lower latency and ultimately, better responsiveness – something that is critically important in numerous applications, including smart lighting controls. Lower duty cycle directly translates into longer battery life, and with Bluetooth’s support for sleepy nodes – devices that spend most of their time in sleep mode, wake up only to quickly perform their task, and then go back to sleep – simple smart “things” (such as sensors) can keep running on tiny coin cells for years.

Other very relevant features of Bluetooth Smart, which make the communication process even more robust and energy-efficient, include low overhead, high spectral efficiency and the adaptive frequency hopping scheme. The latter allows the signal to hop dynamically between the 40 available channels, avoiding the noisy ones and selecting those that ensure a quick and successful delivery. This is particularly important considering the fact that Bluetooth utilizes the same 2.4GHz spectrum as numerous other radio technologies, including Wi-Fi, ZigBee or Thread, but also such appliances as microwave ovens, baby monitors or cordless phones. Still, Bluetooth’s 2.4GHz band is a bit of a downside. While the standard has the tools to effectively counteract interference problems, the 2.4GHz signal weakens faster than a sub-GHz signal as radio waves travel through walls and other obstacles.

That said, the range of the Bluetooth radio is way better than most people think. Bluetooth is still widely considered an ultra-low range communication technology, the one just about sufficient to handle wireless data exchange between a laptop and a cordless mouse. But with Bluetooth Smart, manufacturers may choose to optimize range to ensure a reliable data exchange over incomparably higher distances. Example? After applying certain software and hardware solutions to a standard Bluetooth module, we’ve managed to significantly increase its reach, while ensuring full compliance with the core Bluetooth specification and the RF radiation exposure regulations. Operating at +10dBm Tx power and with -98dBm Rx sensitivity, our modules provide a 108dB link budget that translates to a range of 1500ft (~500m) in the open air. Inside buildings, this value will obviously be much lower and dependent on numerous factors, yet it still remains impressive.

Now if you compare the above characteristics with what other low-bandwidth standards have to offer, you’ll see why we claim that the Bluetooth radio is superior to any other technology available on the market. With a maximum data rate of 250 kbit/s, the single-channel 802.15.4 standard used by both ZigBee and Thread simply can’t deliver such a robustness and resilience. Z-Wave is clearly out of contention, too, with its maximum throughput amounting to a mere 100 kbit/s.

And now the best part. As recently announced by the SIG, the speed of Bluetooth Smart is set to double next year, which means that even lower duty cycle, latency and energy consumption will become achievable. Furthermore, its range will increase up to 4x versus the current standard. Once all these enhancements are implemented, Bluetooth will leave its competitors far behind, especially that none of them has revealed plans of similar improvements regarding their physical infrastructure anytime soon. We’ll have to wait for the adoption of the specification before giving a final judgment, but it does seem like the Bluetooth radio will be stepping up to a much higher weight class in 2016. If everything goes as planned by the SIG, we’ll have a cruiserweight boxer facing lightweight opponents in the IoT communication standards war. Sounds like someone might get hurt in this bout.

But of course it is not all about the radio. We’ve mentioned earlier that Bluetooth is usually not even considered an option in smart home or smart office applications. How could that be if the Bluetooth transport offers such an outstanding set of features? The answer lies in the middle part of the simplified OSI model shown above, with network being the key word.

Despite being developed specifically to address the challenges of the rapidly evolving IoT market, Bluetooth Smart was architected to support relatively simple hub-and-spoke networks. And that’s just not enough to enable a contextual, dynamic, sensor-driven environment that we’d like to see in our connected homes, let alone offices or factories. As we’ve already pointed out in the previous episodes of our “Wireless protocols showdown”, mesh networking is an essential topology for numerous applications, particularly the ones that require extended range or peer-to-peer communication. Smart lighting, and smart building automation in general, are perfect examples. Therefore, it is not surprising that all of the usual suspects in the low-bandwidth category (Z-Wave, ZigBee, Thread) route messages over mesh networks.

Bluetooth doesn’t, at least not officially. In the past, some expressed doubts whether a mesh network can be built on top of the Bluetooth radio, but it certainly is doable. Trust us – we have built our proprietary Bluetooth mesh network from scratch, and it’s been performing outstandingly well. And we are not the only ones. At least several other companies have been developing their mesh solutions over the recent year or so. Transforming a single-hop Bluetooth Smart topology into a robust multi-hop, peer-to-peer network is not an easy process, though. In our case, it took a lot of persistence, creativity and, above all, hard work. But the reward is enormous.

Thanks to the essential features of Bluetooth Smart that we’ve mentioned several paragraphs above, and the ones that we’ll mention below, a Bluetooth mesh network is a powerful and rock-solid backbone for even the most complex building automation networks. Moreover, it performs surprisingly well in certain most challenging applications, such as smart lighting controls. To keep the length of this post reasonable, we won’t get into more details regarding the Bluetooth mesh network here. However, in one of our next blog posts we’ll describe how our proprietary mesh network works to show you what Bluetooth Smart is really capable of.

A very legitimate question one could ask now is: how these proprietary Bluetooth mesh solutions that are being developed by a small group of companies can benefit the market and the user? The answer is, they can’t. We’ve said it multiple times, and we’ll say it again – proprietary solutions aren’t worth much as they won’t solve the problem of interoperability, the IoT’s biggest roadblock to mass adoption. Being fully aware of that, at the beginning of this year the SIG announced the formation of the Bluetooth Smart Mesh Working Group. Its goal is to standardize mesh networking support and incorporate it into the protocol’s core specification. A couple of days ago, the SIG officially confirmed that it’s on track with the development of the Bluetooth mesh, and that the standard will be adopted next year.

Since we brought up the topic of interoperability, let’s see how Bluetooth Smart handles this challenging issue. As shown on the OSI model above, the protocol does cover the application layer, defining profiles for individual applications and types of devices. However, their number is relatively low, and none of them has anything to do with building automation. This is kind of understandable, though, considering that building automation wasn’t previously considered a field where Bluetooth could be of any use. Things are about to change with the release of the mesh standard, so ensuring interoperability at the application layer will become a necessity. A rigorous certification program will also be needed if the SIG doesn’t want to end up where the ZigBee Alliance did (if you don’t know, check out the last 2 paragraphs of our ZigBee blog post).

The SIG has not made any official announcement regarding how it plans to tackle this, so we’ll have to wait until such announcements are made, or perhaps until the new specification is released. What is optimistic, though, is that so far the group has placed great emphasis on ensuring full interoperability between devices bearing the Bluetooth logo. It seems that supporting the application layer has always been a natural direction for the SIG – this is why you can blindly grab any Bluetooth headset off the store shelf without having to wonder whether or not it will work with your phone. You just know it will, and this is the confidence end users should have with every type of smart products on the market. We think it can be fairly safely assumed that the SIG will follow this direction. Anything else would be a shameful waste of an enormous potential that the arrival of the Bluetooth mesh standard is about to unlock.

That’s all about the OSI model, but that’s not all about Bluetooth Smart. While the most popular radio protocols aspiring to connect the IoT have a fair amount of overlapping qualities, there are several value-added features that only Bluetooth can provide. The beacon capability is one of them. Using Bluetooth’s proximity sensing, beacons can make smartphones perform certain actions whenever the user is close to a given smart device. This enables a wide array of unique applications, from location-based push notifications that open up entirely new possibilities in retail marketing, to accurate positioning services that can be used e.g. to guide passengers at the airport to their gates. What’s particularly important is that adding several lines of code to a software stack is all that is needed to equip a Bluetooth-powered smart device with the beacon functionality. Each Bluetooth module is natively equipped with proximity sensing capabilities, so there is no reason why e.g. beacon-bulbs should be more expensive than “standard” smart bulbs.

That proximity sensing is possible due to the fact that a smart device and a smartphone can communicate directly with each other. And this might actually be the biggest single advantage that Bluetooth has over its competitors. Out of all wireless communication technologies used in the IoT, only Bluetooth and Wi-Fi are natively supported by virtually all smartphones, tablets and laptops on the market. But apart from being completely unsuitable for the vast majority of IoT applications, Wi-Fi pushes all messages through a gateway, anyway. Direct communication happens only with Bluetooth. This produces very significant benefits from the user experience perspective, since a phone app is all that is needed to build, configure and control a network of connected devices. Thanks to this direct connectivity, Bluetooth provides the end user with what could be called a Remote Display and a Remote Keyboard for each smart device, no matter how small or simple it is. One of the major benefits resulting from this is the simplicity of the onboarding process. With other protocols, manufacturers often have to come up with the wildest ideas to facilitate adding a new device to the existing network. With Bluetooth, the entire procedure can be much simpler and more intuitive.

But if Bluetooth offers so much goodness, there has to be a price to pay for that, you might think. And you’ll be right. But it’s not about the financial cost of chipsets, as Bluetooth modules are among the cheapest ones on the market. That price comes in the form of complexity. While smartening a device requires the manufacturer to make quite a lot of effort regardless of the technology it decides to embrace, it’s a bit more difficult to implement Bluetooth than to implement one of the other leading wireless communication protocols. On the other hand, that’s quite a universal rule. More sophisticated and powerful solutions are usually more difficult to understand and master.

So this is Bluetooth, a technology in transition. Today, it is still far from being a perfect solution for the connected world, but once the mesh support is standardized and the enhancements announced by the SIG are implemented, it will have everything in place to introduce a new quality to low-bandwidth communications in the IoT. If only the SIG ensures full interoperability at the application layer, manufacturers will need to have a really good reason not to use such a powerful wireless engine in their products.