Riding the Z-Wave
Does the 12-year-old technology still remain a reasonable choice for manufacturers and consumers in the rapidly advancing world of IoT?
Our review continues, and this time we’ll look at one of the flagship home automation technologies of the past decade. Unlike the previously reviewed Wi-Fi, Z-Wave was designed specifically for home automation applications. It’s a low-power wireless communication protocol developed to provide end users with an efficient and reliable method to remotely control a wide range of devices and systems. For manufacturers, Z-Wave offers a cost-effective and easy-to-implement solution for making their products smart and connected.
Z-Wave was introduced to the market in 2003 by Zensys, a company acquired five years later by Sigma Designs which now licenses the technology and remains the primary supplier of Z-Wave chips. Addressing all the most important needs of the emerging smart home segment, it has become the leading international wireless standard for control and automation in residential environment. With more than 1,300 certified devices on the market today, and approximately 35 million compatible units in circulation, it is a mature and proven technology. However, its strong market position is now being challenged by a number of solutions that claim to be better suited for the applications it was intended for.
The impressive market penetration might be Z-Wave’s biggest strength in the race for dominance among wireless communication protocols. For customers, it means the largest selection of interoperable devices for controlling and monitoring their homes. For manufacturers, it creates an opportunity to deliver products that can easily become part of already deployed smart environments, thereby improving their chances of market success. One could expect that the enormous installed base of Z-Wave devices would create a snowball effect enabling the solution to dominate the entire building automation segment, particularly now when the adoption is gaining momentum and the number of connected products on the market is increasing rapidly. However, it must be remembered that technologies come and go, and a strong market presence does not make any solution immortal. What matters in the long run, particularly in the technology industry, is whether a given technology can survive the test of time by adjusting its capabilities to constantly evolving consumer requirements. Twelve years is a long time, especially for such a fledgling market as home automation. A lot has changed since Z-Wave made its debut, so the question whether it still remains a reasonable solution for customers and manufacturers is certainly a valid one.
As mentioned, Z-Wave’s strong market presence benefits customers by allowing them to choose from a wide variety of products when expanding or building their smart homes. They can enjoy this freedom of choice not only because of the exceptionally high number of products on the market, but also because of Z-Wave’s very unique feature, its unmatched cross-vendor interoperability. A quick glance at the simplified OSI model explains it all:
Simple as that, Z-Wave covers all of the layers of the primary reference model for network communications, from the physical layer up to the application layer. We already wrote about the importance of the application layer in the second episode of our series, emphasising its crucial role in overcoming the problem of interoperability in the IoT. Within its own ecosystem, Z-Wave does not have this problem. Unlike the vast majority of leading wireless connectivity solutions, the protocol ensures full interoperability between different branded products based upon it. This wouldn’t be possible without defining the application layer, but also without the strong standard and certification program established by the Z-Wave Alliance. Interoperability is one of the keys to Z-Wave’s market success, especially that the protocol also ensures a complete backwards compatibility with all previous versions, another unique feature considering its relatively long history.
To achieve this, certain sacrifices had to be made along the way, though. A lot of effort was put to ensure that the protocol keeps evolving and its capabilities are enhanced as the time goes by, but the ambitious goal of maintaining backwards compatibility made it impossible to introduce any radical changes. As a result, Z-Wave remains based on concepts that were developed a dozen years ago.
One of these concepts is the mesh network topology. Z-Wave was designed in such a way as to allow individual nodes to forward messages until they reach their ultimate destination. Significantly extending the range of a wireless network, even today mesh networking is viewed as essential for enabling reliable coverage in building automation. When it comes to mesh, Z-Wave is way ahead of numerous other technologies. Bluetooth, for example, is yet to enable such a topology, with a relevant standard expected to be formally adopted next year. Z-Wave has been successfully routing messages over a mesh network for years, although there are certain limitations as to how this communication is handled.
First of all, a Z-Wave network is capable of routing messages via up to 4 repeating nodes. This alone limits potential applications to smart home environment, since commercial premises and office spaces in particular often require a wider network coverage. With a typical indoor range of Z-Wave modules nearing 40m, a mesh network with a 4-hop limit allows for transmitting data over reasonable distances, providing coverage just sufficient for an average apartment, but way below the needs of an entire office building or a vast industrial facility. Furthermore, a Z-Wave network can have up to 232 nodes, although smaller numbers are recommended to prevent saturation. While this seems like more than enough for an average smart home enthusiast, it is, again, a serious barrier for a connected office environment where countless devices, sensors and controllers should work in concert to provide real benefits, particularly with regard to energy management. But even in the residential environment, the above limit might soon turn out to be insufficient. Gartner predicts that within a few years there could be more than 500 smart devices in a typical home, a number that Zensys could not have predicted when designing Z-Wave’s architecture more than a decade ago.
Z-Wave uses source-based routing which means that a device initiating the message generates a complete route through the mesh network to the recipient. Before this can happen, routing tables need to be built. This is done using a device called primary controller which analyses the entire network to come up with optimal routes between its nodes. This source-routed topology is efficient and reliable as long as the network arrangement remains unchanged. But if one of the nodes fails, e.g. a bulb burns out or a mobile device moves out of range, the network’s topology needs to be rediscovered and routing tables must be updated. The healing procedure takes a while (up to 1–2 hours), and the whole network remains down until it’s over. It is recommended to run it regularly during night hours. For some customers this won’t be much of a problem, but the fact that your smart ecosystem cannot be operational 24/7 might be an issue in many scenarios and environments. In the more recent versions of Z-Wave, the so-called explorer frames were introduced that can be used to solve the problem of missing nodes without initiating the lengthy network healing procedure. However, it has been reported that they often cause the entire network to freeze for roughly a minute. From a perspective of the end user not expecting such enormous delays, this could be a highly irritating experience.
In its early days, Z-Wave was widely considered insecure. A number of successful attempts to break the protocol strongly contributed to this reputation, although it must be noted that some of these vulnerabilities were caused by poor implementations rather than Z-Wave’s architecture. Over time, a lot of effort was made to improve the overall security level. As a result, the latest generations of chips feature the government grade data encryption standard AES 128 which requires that a one-time key, commonly referred to as nonce, is sent along with every message.
This does have certain implications, though. Balancing security and convenience is a common challenge for technology vendors these days, but Z-Wave had to pay a particularly high price for ensuring that the exchanged data is protected well enough. While making the entire network significantly more secure and preventing packet replay attacks, the requirement to transport a nonce along with each transmitted message multiplies the communication effort. Before a command can be sent over, a nonce request must be transmitted first, and once it is confirmed by the receiver, a nonce itself can be sent which also requires confirmation from the receiving device. Only after all these steps have taken place, the sender can transmit the message, and the receiver can perform a relevant action. Optimized for exchanging small data packets, typical for home automation applications, Z-Wave offers data rates of 9.6, 40 and 100 kbit/s, depending on the generation of chips. It is therefore very slow, even compared to other low-power communication protocols. After applying all the necessary security measures, which significantly increase the volume of network traffic, end users often have to struggle with latency issues. In some applications this won’t really matter, but there are numerous scenarios where delays go far beyond any acceptable levels. This is one of the reasons why Z-Wave failed to succeed as a reliable solution for lighting control systems where synchronous operation of a set of bulbs requires multiple data exchanges between each of the nodes and the controller.
It is important to realize that these latency issues stem directly from the architecture of the Z-Wave protocol. In general, all of the leading wireless connectivity solutions for the IoT provide a similar level of security. But Z-Wave had to make particularly big sacrifices with regard to efficiency and user experience in order to reach that level.
What Z-Wave does have in common with the vast majority of wireless communication technologies is that a central hub must be deployed within a Z-Wave network in order to enable the end user to control it with a smartphone. This is because the protocol is not directly supported by smartphones or tablets, a feature which remains the exclusive privilege of Bluetooth and Wi-Fi.
One last thing that needs to be mentioned regarding Z-Wave communications model is that the protocol uses an 8-bit checksum with a relatively simple algorithm to verify the integrity of data exchanged between individual nodes. This is often regarded as one of the major weaknesses of this connectivity standard. With an enormous number of data packets running over larger mesh networks, a weak checksum might not be able to prevent an occasional erroneous transmission leading to a totally unexpected behaviour of a smart device. In practice, this means that a light bulb can turn on when not asked to. Or a water valve might open just like that. To address this, a stronger checksum has recently been made available to increase the network’s reliability in some more sensitive applications, but the market still remains full of devices that come with a not-so-exciting pinch of unpredictability.
All of this shows that Z-Wave is certainly not an ideal communication technology for applications that require 100% reliability and efficiency. It also doesn’t seem to be ready to deal with the challenges of the future. Things have gone a long way since it was developed, and it’s now showing some clearly visible signs of aging.
That said, it still remains a perfectly decent solution for smart home enthusiasts who will often be satisfied with their smart home experience despite some of the flaws mentioned above, mainly because their connected environments are relatively small. An average homeowner can enjoy numerous benefits delivered by a smart network of Z-Wave devices, and live through occasional lags or lengthy network maintenance activities. And a random node of a home network performing a random operation once in a blue moon probably won’t cause any major disaster anyway. But for these reasons, we wouldn’t recommend Z-Wave for those more sensitive applications, such as home security, and certainly not for any type of commercial or industrial purposes.
As far as manufacturers of smart devices are concerned, each of them needs to decide whether this is the market they want to target, and whether this is the kind of user experience they want to deliver. An important question to be answered is whether the unmatched market penetration is a good enough incentive to stick to a solution that has its best years already behind it.
Originally published at blog.silvair.com on October 15, 2015.