A Comparison of Common Industrial IoT Protocols

by Tinu Oza, Product Manager, L-com

The Industrial IoT (IIoT) protocol stack can be defined by a number of standards that can, at times, be difficult to sort through. The realm of IoT is notorious for its myriad of competing standards meant to serve overlapping industry verticals. For instance, technologies such as Zigbee^® and Z-wave serve a huge range of smart home use cases to more commercial smart building applications. The same can be said for Bluetooth^® and Low Power Wide Area Networks (LPWANs) like Sigfox or LoRa. Each of these methods of communication has with it its own transport, network, data link and physical layers; each of which are systematically updated. This article provides an overview of the general IIoT architecture as well as insight into the common communication protocols used specifically for IIoT. 

The Smart Factory

According to Million Insights, by 2025 the global IIoT market is expected to reach an astounding value of $922.62 billion. When looking into the general rise of IoT—a forecasted 73 billion IoT connected devices by 2025 (IHS Markit)— the number seems fairly reasonable. Industrial IoT tends to be much more mission-critical, with machine-centric data as opposed to the non-time sensitive, human-centric information in commercial IoT applications. The eventual goal is to feed this data into predictive analytics systems to provide insight into an industrial process with the aim of making the necessary changes at precise times. Industry 4.0 is the term used to describe this process of industrial systems incorporating Cyber-Physical Systems (CPS) for a “smarter” factory. The rise of Industrial Ethernet (IE) feeds into a more seamless implementation of IIoT where central gateways are often directly connected to the cloud with a hardwired Ethernet link, ultimately allowing for a consistent wired backbone across an industrial infrastructure. Back in 2018, IE overtook traditional fieldbus technologies in terms of newly installed nodes for factory automation—an industrial subcategory that requires some of the most demanding latencies and the need for zero downtime on critical equipment. 

Top IIoT Use Case Examples

1. Process improvement and automation
2. Predictive maintenance
3. Asset tracking and monitoring
4. Facility Management
5. Capital Equipment management

These categories can be broken down further into a number of subcategories including process monitoring, process control, and environmental monitoring/control. Some more specific applications can involve visual inspection for quality assurance, sensors/monitoring for machine operation/diagnostics (SCADA), real-time location systems (RTLS) and connected objects for complex edge devices such as inspection drones, industrial wearables, and robot arms. Recent innovations in industrial AR/VR allow for real-time support for technicians. Management of capital equipment may involve stationary machines in a factory, or earth moving equipment and trucks that are mobile.

Industrial Architecture

There is a general three-tier industrial architecture with more commonly used IIoT standards corresponding to specific use cases. This three-tier system goes from the edge, to platform, and finally to enterprise level (See Figure 1).  

Typically at the edge, an industrial wireless sensor network (IWSN) will leverage an array of sensor nodes that interact with a gateway, in a star or mesh topology. The gateway(s) interact(s) with local applications and cloud data aggregators either through a hardwired or wireless connection. Cloud aggregators could be custom proprietary or purchased as a service—IoT Platform-as-a-Service (IoT PaaS)—through a third-party provider, often with user-friendly interfaces and predictive analysis capabilities to prevent faults and downtime. The IoT PaaS collects machine data and makes it much simpler to write applications for end-to-end control. The enterprise level includes the domain applications for human interfacing with rules and controls such as supply chain management (SCM), enterprise resource planning (ERP), and enterprise performance management (EPM). This could also be supplied from a third-party provider as Software-as-a-Service (SaaS). There is an emerging ecosystem of edge-to-enterprise level services for specific IIoT applications.² 

Categories of IoT Nodes

For the purpose of categorizing the myriad of IoT end-device types, the Internet Engineering Task Force (IETF) released a report (RFC 7228) specifying the various classes of “constrained” devices and some strategies for using power for communication (Table 1). The classes depend on the data size and code size of the node. The most severely constrained do not have the resources to communicate directly with the Internet securely as they cannot employ the full protocol stack. The medium-constrained devices cannot easily communicate with Internet nodes, but can use protocol stacks specifically designed for constrained nodes. These classes of nodes also correlate to their power usage where a node leveraging a 802.15.4-based standard would likely fall into the C1/P1 category and a LPWAN node would require far less energy as well as RAM/flash and may best qualify for a C0/P0 category. 

Popular Industrial Automation Protocols

A survey conducted by ON World and International Society of Automation (ISA) every two years scans nearly 200 industrial automation vendors to better understand the various IoT technologies used in their facilities. The most recent survey was released last year and grants some insight into the protocols that are most frequently leveraged.¹ Figure 2 illustrates some of these protocols in comparison to their popularity in 2016; it shows that 802.15.4-based protocols such as WirelessHART and ISA100.11A are used, most often followed closely by WiFi, Bluetooth, LPWANs, and finally cellular connectivity. 

For the sake of brevity, this article will lightly overview and compare the top 5 IoT platforms and their respective IIoT applications excluding other potentially viable options such as cellular (2G/3G, LTE, 5G)  as well as unmentioned networks such as ZigBeePRO or WIA-PA. The IoT parameters of the popular IIoT protocols are listed on Table 2 and may be a helpful reference for the remainder of the article. 

802.15.4-Based (Zigbee, WirelessHART and ISA100.11a)

The standards listed previously serve a particular need of an application depending on the required bandwidth, capacity, reach, and latency requirements. Zigbee, WirelessHART and ISA100.11a are based on IEEE 802.15.4 low-rate wireless personal area networks (LR-WPANs). For these standards, all devices in the vicinity must transmit data periodically to a local gateway that is LAN- or WLAN-enabled. These standards have a maximum operating range of approximately 200 meters and a maximum throughput of 250 kbps. These standards are not specifically designed for a high capacity with a vast range and large number of connected devices, as they have a maximum transmitting power of 10 mW in the 2.4 GHz unlicensed band and are therefore limited in range. This, combined with the requirement need to be in close proximity to a gateway, makes it difficult to deploy tens of thousands of nodes. Note that Zigbee also operates in the industrial, scientific and medical (ISM) radio band at 915MHz in the USA giving it better propagation characteristics. These protocols are, however, designed for long battery life and deep-sleep modes, high network reliability with real-time time-division multiple access (TDMA) transmissions, collision avoidance, frequency diversity, channel assignments, network scheduling, adaptive hopping, and various security protocols. All in all, these protocols are ideal for medium range, medium throughput, and low latency applications, making them a safe choice for many IIoT applications. 

• Benefits: real-time transmissions, reliable, deterministic, low latency
• Considerations: Limited range and capacity
• Applications: Process monitoring, machine diagnostics, environmental, process control 

WiFi

WiFi offers high bandwidth communications that can support more process intensive security protocols. With that being said, there is the issue of wireless coexistence for popular protocols where factories risk external interference with WiFi-dependent devices already in operation. Still, WiFi is often used to augment more application-specific standards (e.g., WirelessHART, sub-GHz protocols) for a more seamless connection to the cloud. There are many multi-protocol IoT modules and SoCs that include Wi-Fi and slower protocols (e.g., LoRa, Bluetooth) for faster OTA upgrades. High data rate use cases such as video streaming with IP cameras for visual production line inspection are well suited for WiFi. One specific use case presented in an Industrial Internet Consortium (IIC) article involves the video detection of hazardous gases in a fire extinguisher factory to adequately warn workers of its presence and location⁴. In cases where feedback and control are tightly coupled, much higher bandwidth and power usage is required to support the frequent uplink transmissions for recurrent updates. Modern Programmable Logic Controllers (PLCs) are also often designed with WAN and WLAN access to support wired and wireless Internet connectivity. Moreover, many IoT installations leverage secure wireless APs to connect to fringe IP devices, effectively increasing the range of IoT installation. 

• Benefits: Large bandwidth (faster OTA upgrades), seamless connectivity to IE infrastructure for device management and predictive modeling  
• Considerations: Interference, power usage
• Applications: Process control, process monitoring, machine diagnostics, asset tracking

Bluetooth Low Energy

Recent Bluetooth standards such as Bluetooth 5 and Bluetooth Low Energy (BLE) offer an enhanced range over older standards (<200m) due to the increase in the maximum transmit power from 10 dBm to 20 dBm. Still, the relatively high data rate of 2 Mbps allows the device to send packets rapidly and more rapidly enter a sleep mode to improve battery lifetime. Bluetooth has the same issue of wireless coexistence as WiFi since it is a prolific technology in the unlicensed 2.4GHz band. Still, Bluetooth is a highly implemented standard in IIoT, likely through its value towards asset tracking applications. BLE beacons and tags offer real-time tracking of factory assets from tools and equipment to shipping containers. This is an improvement over passive RFID technology that requires a user to scan each individual item. And, while the cost of an RFID chip is significantly less than a Bluetooth tag, the infrastructure to support it (e.g., scanners, discreet object tracking) makes RFID less of a viable option over time. ABI research predicts that Bluetooth will account for 52% of the total infrastructure of the $4.5 billion (est.) asset tracking market by the year 2022—the third largest IIoT use case. Some industrial tracking topologies use multi-radio platforms, including a LPWAN protocol with BLE to increase the range of wireless tracking. This combination can also allow for straightforward local device setup and maintenance from existing user equipment (e.g., smartphone) through BLE as opposed to a specialized device for the sub-GHz radio. 

• Benefits: High data rate, localized updates with smartphone/tablet
• Considerations: Interference
• Applications: Asset-tracking, Real-time Location System (RTLS)

IIoT LPWANs 

LPWANs have had notable growth in IIoT applications; the ISA survey cited above noted a two-fold increase from 2016 to 2018. While this technology is still mostly in a “research” phase for most industrial vendors, the top three predicted LPWAN IIoT technologies by 2028 are¹: 

1. NB-IoT
2. LoRaWAN
3. LTE-M

Almost 30% of IIoT vendors involved more than 1000 devices in their IoT installations and over 15% have installations with over 10,000 devices. At capacities beyond 40,000, LPWAN technologies are better suited than the other IoT protocols stated above, so long as they require a low rate of transmissions of small packet sizes. Generally speaking, the LPWAN fills a unique niche of super long range, low power connectivity by utilizing Narrowband (NB) and Ultra-narrowband (UNB) modulation schemes at sub-GHz frequencies. While NB-IoT and LTE-M leverage licensed cellular bands, LoRaWAN uses the 915 MHz ISM band. Each of these protocols are suited to a variety of potential IIoT applications. 

LoRa vs NB-IoT

LoRa and NB-IoT are most similar in regards to ultra-low power, super-long range performance. Still, LoRa is far more optimal in its distance (<20 km) and battery life time. Moreover, LoRa offers a great deal of flexibility with public network operation through base stations and LoRa gateways. However, NB-IoT leverages the existing cellular infrastructure and offers far more data reliability and determinism—the highest priority factors for industrial vendors—at the cost of battery life and range. LoRa is an asynchronous protocol without a clock signal scheduling transmissions between the sender and receiver. On the other hand, NB-IoT is a time-slotted synchronous protocol that offers a more ideal QoS and much lower latencies. NB-IoT is able to send larger data packets per transmission than LoRa, which may be more ideal for more demanding applications. 

With that being said, a very recent development in the LoRa ecosystem is an alliance with WiFi to support more mission-critical applications. A LoRa Alliance whitepaper released in September 2019 summarizes how LoRa can be used on top of an existing WiFi network to benefit from the high data rate, cloud utilization of WiFi and the long range of LoRa³. 

NB-IoT vs LTE-M

In 2017, 3GPP Release 14 made some major revisions to the LTE-M standard that further differentiate these two protocols. Now, LTE-M operates within a relatively massive 5 MHz bandwidth; almost 30 times more spectrum allocated to that of NB-IoT modulation schemes. This allows for a major improvement on maximum throughput of 4 Mbps. The LTE-M1 iteration included a large range of 11 km with a very low latency of up to 15 ms. Similar to NB-IoT, LTE-M1 offers a high level of determinism and reliability with current cellular telephony. The major difference is the throughput, which ultimately requires a trade-off with battery life. LTE-M is also more ideal for mobile use cases, such as fleet management or automated vehicles on shop floors, as it handles seamless handover between cell sites in a manner similar to LTE while NB-IoT would drop the connection and re-establish it in a new cell. A study comparing coverage and capacity of LTE-M and NB-IoT practically analyzed the indoor and outdoor (road users) coverage of the two and found that LTE-M had comparatively excellent coverage for outdoor and road users (99.9%) and light indoor applications (99%), while NB-IoT is able to adequately cover deep-indoor (30 dB penetration loss) applications where far less users (5%) cannot achieve target throughput. Additionally, LTE-M can support 16 times more devices per sector than NB-IoT in indoor applications and up to 1 million (40x more) devices when high-layer security protocols are not used where data can be transferred right after an acknowledgement (ACK)⁴. Table 3 offers a general comparison between these three technologies. Cellular LPWANs can be used for more time sensitive, mission-critical IoT while LoRa offers a more cost-effective alternative for information that does not require a high level of determinism and reliability such as asset-tracking applications with hybrid BLE and LoRa installations. 

Conclusion

There is a tremendous developing ecosystem of protocols for IIoT where many companies are utilizing a hybrid-deployment approach. Over the past four years, the industry has seen major shifts in the use of certain protocols, specifically in regards to the increased use of LPWAN technology to supplement the existing IoT infrastructure. Each protocol has its own benefits, considerations, and particular IIoT applications it may be best suited towards.  

References:

1. https://www.automation.com/pdf_articles/ON_world/InTechMagazineInsert_ONWorldFinal.pdf

2. https://www.iiconsortium.org/news/joi-articles/2017-Jan-Smart-Factories-and-Challenges-of-Proximity-Network.pdf

3. https://lora-alliance.org/resource-hub/wi-fi-lorawanr-deployment-synergies

4. M. Lauridsen, I. Z. Kovacs, P. Mogensen, M. Sorensen and S. Holst, “Coverage and Capacity Analysis of LTE-M and NB-IoT in a Rural Area,” 2016 IEEE 84th Vehicular Technology Conference (VTC-Fall), Montreal, QC, 2016, pp. 1-5.

5. Mekki, Kais, et al. “A Comparative Study of LPWAN Technologies for Large-Scale IoT Deployment.” ICT Express, vol. 5, no. 1, 2019, pp. 1–7., doi:10.1016/j.icte.2017.12.005.

6. E. Sisinni, A. Saifullah, S. Han, U. Jennehag and M. Gidlund, “Industrial Internet of Things: Challenges, Opportunities, and Directions,” in IEEE Transactions on Industrial Informatics, vol. 14, no. 11, pp. 4724-4734, Nov. 2018.

(3948)