IoT Protocols: A Growing Stack

Someday there may be a single protocol used for all IoT applications, but don’t bet on it. More and more solutions are emerging such as the Thread protocol has the backing in the formidable resources of Google.

You’re probably tired of hearing, reading, and seeing the letters “IoT”, the seemingly inescapable “Next Big Thing” that is about to change our lives forever. The key word is “Next”, as most of the applications it will enable have been a long time in coming, and are still coming. It’s not for a lack of technology but too much of it, as incompatible protocols continue to proliferate with no clear winner in sight. Considering the projected immensity of the IoT market, such competition is hardly surprising, and it’s not unprecedented (remember the Betamax-VHS war?) but this time there are not two but perhaps a dozen protocols and consortia vying for supremacy. To this list, we must now add Thread.

The Thread protocol (http://www.threadgroup.org/) owes its existence to “smart” thermostat and smoke/carbon monoxide alarm manufacturer Nest Labs, which was acquired last July by Google for $3.2 billion. A month earlier Nest had announced that it was establishing a working group to create an IoT protocol with silicon vendors ARM Holdings, Freescale Semiconductor, Samsung, and Silicon Labs as partners, along with lock maker Yale and Big Ass Solutions (formerly Big Ass Fans, which makes…well you get it).

Since then, more than 80 companies have jumped on board either as partners or affiliates, which is typical of promising new protocols, as every major competitor has at least this many supporters and every company wants to hedge its bets. A version of Thread has actually been integrated in the Nest Thermostat for some time, but not enabled.

Thread has several strengths that should make it a formidable competitor to ZigBee, Z-Wave, and Bluetooth Low Energy (BLE), also known as Bluetooth 4.1 and Bluetooth Smart, its strongest competitors. First, Nest/Google adopts the path taken by ZigBee and Z-Wave rather than creating something entirely new, basing Thread on the well-established IEEE 802.15.4 wireless standard and IPv6, the latest version of the Internet Protocol (IP). IPv6 extends the number of possible IP addresses to an unimaginably high 2^128, taking over from its predecessor IPv4 that topped out at 4.3 billion, a number that was reached in 2011. The Thread architecture is shown in Figure 1.

Figure 1: The main components of the Thread protocol. Thread targets appliances, access and climate control, energy management, lighting, safety, and security. (Source:http://www.threadgroup.org/Portals/0/documents/events/ThreadIntro.pdf)

Second, while administered within Nest, Thread is championed by Google, giving it massive resources, market clout – and money. Shortly before it was acquired, Nest bought Dropcam, a privately-held manufacturer of innovative, easy-to-use Wi-Fi-based webcams and trackers that lets its users check on their connected cameras from virtually anywhere, and stores files of captured video in the cloud for later inspection. Given Google’s almost limitless resources for data collection and geolocation, combined with the Thread protocol, Dropcam’s cloud-based video surveillance capability, Android-based smartphones, and Nest’s information about its user’s homes and habits makes for a powerful combination. It also raises the eyebrows of critics who point out that little is left that Google cannot find out about us (a criticism that has also been leveled at IoT in general).

It would seem that the most vulnerable target for Thread is BLE. Bluetooth may be ubiquitous but it never intended to be used for personal networks and its specifications reflect this. That, as we’ll see, is changing fast. The most significant drawback of BLE has been that it cannot form a mesh network of self-healing devices that reroute traffic if a device or link on the network fails, keeping the system alive. Mesh networks are essential for IoT applications like home automation, which include thermostats and security alarms that must be extraordinarily reliable. This is equally true of industrial networks and even more so for networks operating in a public safety or other mission-critical environment, and for connecting massive numbers of devices within a city, to name just a few examples.

Another major inherent disadvantage of BLE for use in IoT is its maximum range, which is limited to about 330 feet but in practice is typically much less. In a home automation or other network environment of a reasonable size this is obviously a significant problem for protocol without mesh networking capability. In contrast, IoT solutions based on IEEE 802.15.4 can operate over a range of about 100 feet, which can be massively scaled upward by connecting networks.

Since Thread, ZigBee and others (but not BLE) support the 128-b IPv6 addressing scheme, they are more or less future-proof in terms of IP addressing, which is mandatory as IoT devices proliferate. However, although BLE currently does not support IPv6, the next generation is likely to.

Is Bluetooth DOA for IOT?

While the rest of the IoT standards groups and industry consortia race to solidify their place and finalize their standards-based or open-source strategies, the Bluetooth Special Interest Group, Broadcom, Qualcomm, and other companies have been working to eliminate the current limitations of this standard for IoT applications.

BLE may have limitations, but it also has strengths, and is sprouting new capabilities beyond those of Classic Bluetooth and even BLE, which was introduced in 2010. For example, Bluetooth 4.1 introduced in 2013 has a data rate of 1 Mb/s, which is faster than many IoT-centric competitors other than Wi-Fi, but also slower than “Classic” Bluetooth. Bluetooth 4.1 has standard 128-b AES encryption and reduced latency from 100 ms to 6 ms or less.

A good example of a highly-integrated BLE SoC is Broadcom’s BCM20737 WICED SMART Bluetooth device that has security features and iBeacon technology using its low-power WICED Smart chip. It includes RSA 4000-bit encryption and decryption support and includes native support for A4WP Rezence wireless charging.

Figure 2: A good example of a highly-integrated BLE SoC is Broadcom’s BCM20737 WICED SMART Bluetooth device.

The Bluetooth SIG is also paving the way for connection to the Internet, and in February launched the Bluetooth Smart Mesh Working Group (already supported by more than 80 companies) that has the goal of building the architecture for standardized Bluetooth mesh networking capability. Once the specification is released, the group has eliminated the standard’s greatest drawback for using it in IoT applications: the lack of a mesh network architecture. This effectively moves Bluetooth back into the top tier of contenders, which is an impressive feat for a technology that in 2003 was called by a contributor to EE Times as …”toast, finished, over. Stick a fork in it. It's done.”

Group Think

IPv6, IEEE 802.15.4, and a personal area network called IPv6 over Low-power Wireless Personal Area Networks (6LoWPAN) used by Thread, ZigBee, and Z-Wave are complementary, as the latter two were conceived expressly to serve devices with limited processing power, low data rates, very low RF output power, and minimal power consumption from either mains or battery. This should make device and network design relatively straightforward and cost-effective.

All use symmetric-key 128-b AES encryption that has 3.4x1038 possible combinations and would take a current supercomputer about 149 trillion years to break. It is considered by the National Security Agency as safe enough for security up to SECRET level. The need for such security has taken on increased importance after a company called Context Information Security Ltd. of London announced that it had hacked into a smartphone-controlled network of lightbulbs. They used 6LoWPAN and Wi-Fi as their wireless transmission scheme. The lightbulb breach generated an enormous amount of bad publicity for the company. The need for IoT security was lost on no one, as Thread’s web site and others make abundantly clear.

One device family for ZigBee is NXP’s JN5168-001-Mxx family of low-power modules for IEEE 802.15.4, JenNet-IP, ZigBee Light Link, and ZigBee Smart Energy that use NXP’s JN5168 wireless microcontroller and has all RF components onboard. It complements wireless control or sensing products that together with a power supply and peripherals such as switches, actuators and sensors, can form a complete product.

Figure 3: NXP’s JN5168-001-Mxx

Thread’s other features include the ability to accommodate between 250 and 300 devices on a network, and to communicate between devices or between devices and the cloud so authenticated users and their mobile and fixed devices have access to the network from any Web-based connection. Like its competitors, Thread defines only the physical and MAC layers while leaving the upper layers open for application development, has low latency of about 100 ms that is less than Wi-Fi-based solutions, and lets any device react almost instantly.

In summary, Thread offers the essential benefit of mesh networking, a cloud-based approach, unlimited IP address availability via IPv6, 128-b AES security, the range required for large networks via the IEEE 802.15.4 standard, the IoT-centric routing benefits of 6LoWPAN, low latency, and broad flexibility for application development, among other attributes. Collectively, these attributes make Thread competitive with ZigBee, Z-Wave, and the rejuvenated BLE. Of course, it also has Google’s undeniable clout and resources to move it along.

All of these strengths don’t mean that Thread will trounce competitors to become the standard for all of IoT. Although IoT has sprouted in a variety of applications, these are early days and many issues remain to be resolved. The most onerous of these is how long it will take before a single protocol emerges as a victor, if in fact one ever does. The best vision of IoT depends on sharing a standard protocol. It would hardly be the first time that competing interests have fought to the death before one reigns supreme. However, IoT is projected to be large enough to support various solutions that might be deemed better suited for some applications than others. The anticipated growth in IoT is shown in Figure 4.

Figure 4: Projected growth in wireless-enabled devices showing the exponential growth of IoT. (Source: Business Insider, http://www.slideshare.net/bi_intelligence/bii-the-internet-of-everything-2015)

The Competitors

To get a better feeling for where Thread “fits” it helps to provide some details about the competition. Besides those mentioned below, there are other groups in the US, Europe, and elsewhere that are also developing their own strategies for IoT, but those mentioned here are a reasonable representation of the top-tier.

ZigBee 3.0: Highly-developed mesh network standard operating at 2.4 GHz and supported by 400 manufacturers. Its maximum data rate is 250 Kb/s, is scalable to support thousands of nodes over a distance of about 100 feet (defined by IEEE 802.15.4 standard), and supports IPv6 and 128-b AES encryption. The “3.0” version, available to members now and expected to be ratified in the third quarter of 2015, integrates all the formerly discrete application-specific versions of ZigBee including ZigBee Home Automation, ZigBee Light Link, ZigBee Building Automation, ZigBee Retail Services, ZigBee Health Care, and ZigBee Telecommunication. The proliferation of these ZigBee variants has been widely criticized by its competitors, which this version will solve. ZigBee Alliance: www.zigbee.org.

ZigBee Pro: New mesh network variant of ZigBee that operates in the 900 MHz unlicensed bands as well as at 2.4 GHz, can accommodate up 64,000 nodes per network and many times that by linking networks. It is frequency-agile over 16 channels, supports multiple star topologies and inter-personal area networks, spread-spectrum modulation techniques to avoid interference, and various transmission options including broadcast. It has a ZigBee Green Power option that allows devices without batteries to join the network, making energy-harvesting techniques such as motion, light, and vibration for extremely low-power operation sources of power. ZigBee Alliance: www.zigbee.org

Z-Wave: Well-developed mesh network standard operating around 900 MHz (varying by country) that is supported by more than 300 companies, and is designed to provide product interoperability regardless of application, rather than application-specific variants like ZigBee (before Version 3.0.) Its maximum data rate is 100 kb/s and it can support up to 232 nodes over a distance of about 100 feet between nodes. It incorporates IPv6 and 128-b AES encryption. Z-Wave Alliance: www.z-wavealliance.org

AllJoyn: Emerging open-source, collaborative software framework allowing developers to write applications for IoT regardless of brand, category, transport mediums, and operating systems without the need to use the cloud or even the Internet (both of which are supported, however). It provides support for Wi-Fi, Ethernet, serial, and power line transmission media (and soon others). Supported operating systems include RTOS, Arduino, Linux, Android, iOS, Windows, and Mac. The framework uses 128-b AES encryption and is currently supported by more than 120 companies. Allseen Alliance: www.allseenalliance.org

CSRmesh™: Created by CSR (formerly Cambridge Silicon Radio), it allows an almost unlimited number of BLE-enabled devices to be networked together and controlled from a smartphone, tablet or PC, effectively creating a Bluetooth network for the first time. www.csr.com

Open Interconnect Consortium: Open-source industry consortium now under the umbrella of the Linux Foundation focused on improving interoperability and defining the connectivity requirements for IoT. It uses a common communications framework to wirelessly connect and manage the flow of information among personal computing and emerging IoT devices, regardless of form factor, operating system or service provider. The consortium recently launched a preview release of its IoTivity specification. www.iotivity.org

Will Chaos Reign?

Keeping track of all the players in the rapidly-emerging IoT market is confusing and complicated by the fact that it seems to change by the week. It seems likely that BLE, ZigBee, Thread, and Z-Wave will continue to compete for a long time, as each one has its own strengths and supporters. That said, once BLE gains mesh networking support within the Bluetooth SIG, all four standards above will have the necessary accoutrements for IoT, making support for a particular standard a matter of choice rather than technological superiority. Nevertheless, there will still be room for open-source versions that might be very appealing for manufacturers looking to create a compatible and interoperable, although differentiated, approach.

Barry Manz is president of Manz Communications, Inc., a technical media relations agency he founded in 1987. He has since worked with more than 100 companies in the RF and microwave, defense, test and measurement, semiconductor, embedded systems, lightwave, and other markets. Barry writes articles for print and online trade publications, as well as white papers, application notes, symposium papers, technical references guides, and Web content. He is also a contributing editor for the Journal of Electronic Defense, editor of Military Microwave Digest, co-founder of MilCOTS Digest magazine, and was editor in chief of Microwaves & RF magazine.