Network infrastructure: your future business foundation – Part 2
Taking advantage of performance enhancements from new technologies
In Part 1 of this article the various aspects of migrating to, and managing, an industrial Ethernet infrastructure were discussed. And like any other critical business and plant infrastructure, it is important to look forward to emerging technologies in order to prepare for future technologies that may further enhance performace, reliability and productivity.
Network planning next two years
Networks continue to evolve over time. Plotting the evolutionary path for the business network requires insight into future trends and possibilities. This section offers insight via identification and discussion of important technologies. The time domain for these technologies is within the next two years.
Power over Ethernet
Power over Ethernet (PoE) is an Ethernet-compatible technology created to enable Voice over IP (VoIP) telephony. DC power, at a nominal voltage of 48 VDC, is carried on one or more pairs in the Ethernet cable along with the transmitted signal. PoE-powered devices (PD) negotiate with the power source (typically a network switch) to ensure appropriate power is delivered. Businesses soon realised the potential of network supplied power. PoE now powers IP cameras, wireless access points, badge readers and access gateways, and office lighting (Figure 1).
Today, PoE in industrial networks performs identical tasks to those it performs in enterprise networks today. These tasks include powering shop floor phones, wireless access points and IP cameras. PoE holds a bright future as the standards community expands its capabilities.
PoE is a key technology for the future of industrial networks because, with the advent of IEEE 802.3bt, conspicuous amounts of power can be delivered along the Ethernet connection to a device. With 71 W available at the end of an Ethernet cable, device manufacturers can be very creative. This ‘one wire ideal’ allows device power and communications in a single connection, simplifying all phases of the device lifecycle. In doing so, PoE alters the DC power infrastructure of control systems.
Legacy protocols are serial communications to the device at a very modest data rate. No device power is delivered by the connection. Therefore, local DC power supplies are required near the device to meet its power requirements. Behind the DC power supply are many AC components (eg, transformer, connection wires, circuit protection, etc) to convert machine mains power to a usable input to the DC power supply. When this supporting infrastructure can be eliminated, control system DC power infrastructure is simplified and costs become lower.
PoE negotiates with the device at start-up to determine the appropriate power level to deliver. There is no need for preconfiguration of each circuit in a standards-compliant installation. Additionally, since device power is controlled by PoE-enabled ports in the switch above it, toggling device power can be done via network switch commands, simplifying service procedures.
PoE should figure prominently in new network installation to simplify powering needed by devices like cameras and wireless access points. The transformative effect on the DC power infrastructure, while quite feasible, is going to take longer to become a reality.
Single pair Ethernet
Work is underway in IEEE 802.3 to create standards for Single pair Ethernet (SPE). Many variants are proposed from short reach (15 to 40 m at 1 Gbps transmission speed) to extreme lengths (up to a kilometre at 10 Mbps transmission speed), all over a shielded twisted pair cable. For industrial network applications, the variant to watch is IEEE 802.3cg, the 1 km at 10 Mbps variant. All variants of SPE are considering a methodology for power delivery like PoE called Power over Data Line (PoDL), IEEE P802.3bu.
Single pair Ethernet drives ‘Ethernet-to-the-edge’ and is a vital portion of legacy protocol migration plans. For 802.3cg, its 10 Mbps transmission speed provides more than enough bandwidth for end device and sensor data rates. For industrial networks:
- The reach objective of up to 1 km is ideal for plants with a large footprint, (eg, oil and gas, petrochemical, etc).
- Power delivery via the Ethernet connection achieves the aforementioned ‘one wire ideal’ where device power and communications are enabled by a single connection.
- SPE, being standard, unmodified Ethernet, supports purpose-built Ethernet forms like EtherNet/IP and Profinet without issue.
- Conductor wire gauge for SPE will need to be at least 18 AWG to achieve the 1 km reach objective; a latent benefit of this cable construction is the ability to drive higher current levels than 4-pair cable, making LED lighting installations more effective.
- SPE media should be easily field terminable; this aspect can reduce pre-terminated cordset inventories and address slack management issues.
Single pair Ethernet will rise to prominence by taking Ethernet to the edge of industrial networks. Device manufacturers and network switch manufacturers are closely monitoring and contributing to the creation of IEEE standards that enable this future concept.
Power over Data Line
Power over Data Line (PoDL) is governed by IEEE standard 802.3bu. The PoDL acronym is frequently pronounced “poodle” in conversation. It represents a necessary adaptation of PoE. A reasonable question is “why can’t we just use PoE on SPE?” The reason is PoE requires at least two pairs to work. This is because there is an electrical connection between pair centre taps (Figure 2a).
Since SPE has only one pair, the PoE circuit (above) does not work. However, a simpler circuit with a lowpass/highpass band splitting filter network works with SPE (Figure 2b).
Using PoDL Class 8 and Class 9, PD power can be 30 W or 50 W respectively at 100 m. New classes are required to accommodate the expected higher loop resistance of 1000 m links seen in 802.3cg.
PoDL and SPE go hand in hand as technologies to watch and include in legacy protocol migration plans.
Wireless sensor networks
Wireless sensor networks are gaining popularity as businesses seek solutions that improve decision speed and quality. Wireless networks can be implemented quickly, speeding the availability of additional knowledge to achieve these goals.
Speed and reliability, however, are not yet at the same level as wired connections. Critical control connections will remain wired for the foreseeable future. Nevertheless, wireless connections provide a fast and cost-effective means to collect additional data to propel analytics efforts, study new facets of an existing process, etc.
There are many wireless sensor networks worthy of consideration but two stand out for industrial applications. These are:
- Wireless Mesh
Most wireless mesh networks used for wireless sensor applications are based on IEEE 802.15.4. This is the technical standard which defines the operation of low-rate wireless personal area networks (LR-WPANs). Wireless mesh networks have a unique feature that make them a provocative choice for industrial data collection; they are self-healing. If a wireless sensing node is blocked from communicating directly with the sensor gateway, it will ‘hop’ to an adjacent node to get back to the gateway. This feature is superb for industrial applications given the dynamic environment with material handling equipment and other large metallic structures often in motion. For example, the chances of a forklift mast blocking transmission at some point in the day is easily an even odds bet.
Low-power wide-area network (LPWAN) is another technology worthy of note for industrial applications. LPWAN protocols are well suited for use in industrial settings. These networks are nominally 900 MHz, a frequency range that performs well in highly metallic environments.
LoRaWAN is intended for wireless battery-operated nodes in a regional, national or global network. It targets key requirements needed for the Internet of Things (IoT) like low data rate, low cost and long battery life while delivering vital features such as secure bidirectional communication, location and mobility services.
End devices using LoRaWAN can choose from three device classes, allowing different device behaviour depending upon optimisation needs:
- Class A — battery-powered node. Class A operation optimises communications to conserve battery power at the node.
- Class B — low latency needed. Class B devices open extra receive windows at scheduled times to optimise communications but with shorter battery lifespan.
- Class C — no latency. Class C devices have nearly continuously open receive windows reducing latency to its practical minimum.
More information can be found at www.LoRa-Alliance.org.
LoRaWAN presents performance advantages for wireless sensing networks in industrial environments and is already gaining popularity for many IoT applications. This is a wireless network to watch and include in your future network planning.
Network planning beyond two years
Industrial network evolution includes a strong influence of better IT/OT collaboration. As these two very capable groups act in concert to improve business outcomes, some advanced IT practices will find their way into industrial networks. One of these is time-sensitive networking (TSN).
TSN gets a lot of attention from automation experts due in part to the increased interest in the Industrial Internet of Things (IIoT). Some of the data collected by IIoT sensor networks is not inherently time sensitive. However, some data is mission critical and time sensitive and must be shared with strict latency and reliability requirements. Further, all data is enriched by adding accurate time context as it allows correlation and analytics to excel. Therefore, TSN is an important technology both within the control loop and outside the loop in IIoT applications.
There are four key benefits that TSN applied to industrial networking provides:
- Bandwidth: Machine vision, 3D scanning and power analysis applications running on an industrial network create large datasets which can strain available bandwidth, but proprietary Ethernet derivatives that are used in industrial control today are limited to 100 Mb of bandwidth and half-duplex communication. TSN supports standard Ethernet in full duplex with higher bandwidth options such as 1 Gb, 10 Gb and even a projected 400 Gb version.
- Security: TSN embraces top-tier Ethernet security provisions: segmentation, performance protection and temporal composability add multiple levels of defence to the security framework.
- Interoperability: TSN integrates existing brownfield applications and standard IT traffic by using standard Ethernet components. It inherits many existing Ethernet features like HTTP interfaces and web services, and these features enable remote diagnostics, remote visualisation and repair capabilities common to IIoT systems.
Latency and synchronisation: TSN prioritises low-latency communications to provide fast response and closed loop control applications. It can achieve deterministic transfer times in the order of tens of microseconds and time synchronisation between nodes down to tens of nanoseconds. TSN provides automated configurations for high reliability data paths where packets are duplicated and merged to provide lossless path redundancy, and in doing so, it ensures reliable delivery of time-sensitive traffic.
TSN provides network designers with tools to ensure that critical traffic is received in a timely and reliable manner. It also frees up congestion to allow non-critical traffic to be converged onto the network and move as ‘best effort’ traffic. This is an essential distinction in that almost all traffic is best effort. Wire speed and limiting traffic to only critical message streams is used to make the network function correctly.
There are two groups to monitor regarding TSN. These are:
IEEE 802 has united several domain experts under the auspices to 802.1 to create a suite of TSN specifications that are without equal. The group has led application needs from audio/visual, automotive, industrial automation and consumer realms in creating these specifications.
The AVnu Alliance focuses on the creation of an interoperable ecosystem through solution certification. Member companies include National Instruments, Broadcom, Cisco and Intel, to name but a few. The AVnu Alliance website presents superb resources to help companies understand and adopt these powerful concepts.
The industrial network infrastructure is a valuable business asset. Investments in legacy industrial networks require a clear migration path to optimise return on assets while not missing out on performance enhancements from new technologies. A robust, well-executed physical layer is foundational to this asset continuing to deliver value. Rapidly emerging technology advances such as the Internet of Things, wireless sensor networks, Power over Ethernet and time-sensitive networking can further leverage your network with a little education and planning.
Correctly planning and managing this vital business asset is crucial to future business success.
The benefits to automation of the latest standardisation efforts.
Modern industrial wireless systems overcome the traditional problems with wireless and mark a...