“The IoT revolution is just around the corner, and by 2020 there will be more than 30 billion connected things in the world. At a time when the world’s population continues to grow and resources become more valuable, this interconnection promises to provide real-world data to drive efficiencies and simplify business practices.
The IoT revolution is just around the corner, and by 2020 there will be more than 30 billion connected things in the world. At a time when the world’s population continues to grow and resources become more valuable, this interconnection promises to provide real-world data to drive efficiencies and simplify business practices.
With the widespread adoption of the Internet Protocol (IP), it has become easier to process data and make the most of information. Fortune 500 companies provide enterprise and database solutions for data storage and software tools to simplify business processes such as asset tracking, process control systems, and building management systems (see Figure 1). Smartphones and tablets provide people with very useful information, such as providing real-time parking information, or real-time monitoring information about the health of machines to make maintenance plans. Many wireless sensors have been deployed to date, but more sensors are still urgently needed to provide data in order to measure and optimize processes that have not been involved before.
Figure 1: Reliable, Low-Power IP Wireless Sensor Networks Will Be Widely Used
To further scale sensor deployments, work is underway on IP standards with the goal of making small wireless sensors as easy to use as web servers. There are two driving forces behind the standards development work: one is the well-recognized high reliability performance of low-power, time-synchronized mesh networks; Seamless integration of the Internet. The combination of these two forces will effectively facilitate the reliable communication of small, low-power sensors, and eventually enable such sensors to be IP-enabled.
Challenges posed by wireless sensor networks
Wireless is inherently unreliable, so it is important to understand the sources of unreliability in order to address unreliability in communication systems. In low-power wireless networks, the main sources of unreliability are external interference and multipath fading. Interference occurs when an external signal, such as a WiFi signal, temporarily prevents two radio systems from communicating. This requires them to retransmit the signal, thus consuming more power. Multipath fading occurs when a wireless signal bounces off an object in a nearby transmitter, and the various echoes can cause destructive interference to the receiver antenna. This phenomenon is a function of the location of the device, the frequency used, and the surrounding environment. Because the environment surrounding any wireless system changes over time, any RF channel will experience problems during the operational lifetime of the wireless system. However, multi-pass fading is affected by frequency. So while a certain frequency may be experiencing problems, there are still other RF channels that are working properly. Because of interference and multipath fading, the key to building reliable wireless systems is to diversify channels and paths without sacrificing the benefits of low-power operation. Dust Networks (now a business unit of Linear Technology) was the first to offer such a system using time-synchronized, channel-hopping mesh networking technology.
Time-synchronized, channel-hopping mesh network
In a time-synchronized channel-hopping mesh network, all wireless nodes on a multi-hop network are synchronized to within tens of microseconds, and the time is divided into time slots. Communication is coordinated through a schedule that dictates what each node should do in each time slot (transmit, receive, sleep). Because they are all synchronized, these nodes turn on their radios only when communicating, greatly reducing the radio duty cycle (1% is common) and extending battery life. In addition, because the schedule can be set flexibly, the network is always available to the application, unlike other “sleep” network architectures that require a complete shutdown of the network for extended periods of time. All packets sent between two nodes are transmitted on frequencies calculated in a pseudo-random frequency hopping fashion. The resulting frequency diversification is an effective way to counteract interference and multipath fading. Time-synchronized mesh networks deliver up to 10 years of battery life and >99.999% end-to-end reliability.
Time-synchronized mesh network a success
In recent years, time-synchronized channel-hopping mesh networks have been widely used. In 2004, Dust Networks first introduced the SmartMesh® system, and industrial processes were one of the first to adopt the system.
Some industrial applications operate in the harshest environment and have the most stringent requirements for data integrity. If such high data integrity can be guaranteed, the efficiency, productivity and safety of industrial equipment can be greatly improved. Because traditional wired industrial sensors are expensive to install, measurements are generally made at only a small number of possible measuring points in the factory. Although this has created a great demand for the use of wireless sensors in industrial applications, traditional point-to-point wireless systems lack the required reliability and are difficult to install, limiting the use of wireless systems in small and isolated applications.
With the introduction of time-synchronized mesh networks, wireless systems can provide reliability typically only available with wired systems, so the application of low-power wireless systems becomes a reality. Low-power wireless systems are standardized through the industry standard IEC62591 (also known as WirelessHART), enabling interoperability of devices in the industrial process market. Most large industrial manufacturers such as Emerson Process, Siemens, ABB, Endress Hauser, Pepperl Fuchs and Phoenix Contact are delivering Wireless HART products. Today, SmartMesh networks are widely used, with more than 30,000 networks deployed in more than 120 countries around the world, improving safety and efficiency in a variety of locations, including steel mills and refineries, remote oil fields and offshore drilling platforms, and Food and Beverage Factory (Note 1).
In addition to industrial processes, SmartMesh systems have also been successfully deployed in data centers and commercial office buildings to optimize air conditioning costs (Note 2). Streetline Networks (Note 3) is a smart parking service provider that monitors the availability of parking spaces in urban areas in real time (see Figure 2). The vehicle detector is installed under the parking space, in the road surface of the driveway. This poses challenges because the antennas of the sensor equipment are located underground and, when the parking space is occupied, are covered by the metal car body. Such applications, previously considered impossible or impractical, can now be achieved with time-synchronized channel-hopping mesh networks.
Figure 2: Streetline Networks deploys a time-synchronized channel-hopping network to improve parking conditions in urban areas like Hollywood, California, USA.
In network technology, standards play an important role as end users embrace solutions developed based on standards. When a technology is developed and approved by a major standardization organization, users can use it with confidence. WirelessHART / IEC62591 is a standard in the industrial process field, and outside this market, Internet Protocol (IP) is a communication standard.
All devices connected to the Internet communicate with each other using IP. Each device gets an IP address, which unmistakably represents that device on the Internet. The packets exchanged contain an IP header and a series of bytes (the address code of the device that created the packet and the address code of the destination device). There are many other protocols (TCP, HTTP, etc.) that are required to form a protocol stack, but the IP protocol is the common denominator. The use of IP protocols to connect low-power mesh network devices to the Internet has made a major contribution to the development of the Internet of Things.
Several standardization organizations have developed standards for the Internet of Things (see Figure 3). The challenge is to achieve full integration with the Internet while incorporating the proven technology of time-synchronized channel-hopping mesh networks. Most of the protocols used on the Internet today are developed by the Internet Engineering Task Force (IETF), and the CoRE working group of this standardization organization has developed the CoAP (Constrained Application Protocol) at the application layer. CoAP runs on top of the UDP protocol and is easily converted to HTTP to enable network-like interaction of wireless sensor nodes. The 6LoWPAN working group has developed an IP adaptation layer protocol that compresses the large headers of IP packets into small wireless frames or packets, enabling sensor nodes to be individually addressed by IP address. Although these upper-layer protocols enable network-like interaction and integration with the Internet, it is the protocol layers below these upper-layer protocols that determine the quality of WSN communication.
Figure 3: IP stack for low-power, reliable wireless sensor networks
The standards formulated by the IETF generally run on wireless chips that follow the IEEE802.15.4 standard. IEEE802.15.4 makes a healthy trade-off between data transfer rate (250kbps), range (10 to 100 meters), power consumption (5mA to 20mA when sending or receiving), and packet size (up to 127 bytes). This trade-off makes IEEE802.15.4 very suitable for low-power mesh network technology, so the standard has become the de facto standard for this type of network link technology.
In 2012, the IEEE published a new medium access standard, IEEE802.15.4e, that runs on IEEE802.15.4 compliant radios. Its Time Slot Channel Hopping (TSCH) mode incorporates Dust Networks’ time synchronization mesh protocol for precise time slot synchronization and RF channel hopping.
Although IEEE802.15.4e defines a mechanism for two nodes to achieve synchronous packet transmission, it does not define how to assign a schedule to each node. This communication schedule enables the TSCH network to flexibly match the communication needs between network nodes (see Figure 4). For example, a network can be configured as a small network with low data transfer rates and very low power consumption, as is commonly found in remote environmental monitoring applications. The same network can also be configured as a large network optimized for faster data transfer rates. In addition, the automatically assigned yet flexible schedule allows the TSCH network to adapt to the surrounding environment. In particular, scheduling enables network functions such as self-healing, route optimization, and load balancing, which are critical to delivering high performance over the network lifetime. Solutions for establishing and allocating TSCH schedules can be developed, but interoperability of wireless transmissions is not possible with such solutions until the standard is in place.
Figure 4: Small TSCH-based wireless nodes such as Linear Technology’s LTP5901-IPM provide 5 to 10 years of battery life with >99.999% data reliability.
However, the above situation has changed with the development of a new standardization work at the IETF, also known as 6TSCH (Deterministic IPv6 over IEEE802.15.4e Time Slotted Channel Hopping – Note 4). This standardization effort, jointly led by Linear Technology and Cisco Systems, will develop the currently missing communication protocol to allow TSCH schedules to be managed by scheduling entities.
6TSCH fills the remaining gaps in the IP protocol stack and will enable fully standardized, interoperable, IP-based wireless sensor networks, providing the high reliability typically only available in wired sensor networks. Network developers will be able to obtain sensor data in real-time by sending network requests to the sensor’s IP address, and the underlying wireless sensor network will support this type of communication with >99.999% data reliability. By making sensors as accessible as web servers, wireless sensor networks will be able to feed the IoT with real-world information.