By 2023, the number of Industrial Internet of Things (IIoT) active connections is predicted to surpass 46 billion, representing a staggering growth of 140% over the next four years. IIoT applications like remote monitoring and predictive maintenance require massive sensor networks that can continously capture granular data from assets and their functioning environments. With the explosion of data acquisition points and traffic at the edge, industrial companies are in critical need of a scalable machine-to-machine (M2M) communications architecture.
Engineering a scalable solution from the beginning minimizes the need for expensive modifications in the future, thereby securing long-term return-on-investment (ROI). However, building such an architecture does not come without challenges. There are many factors to take into consideration when choosing the right “last-mile” connectivity.
6 Considerations for designing scalable IIoT communications solutions
1. Network capacity
Network capacity refers to the amount of traffic a network can handle at a given time. High network capacity allows for seamless and reliable system operations as the number of end nodes continuously grows.
Self-interference caused by confined radio resources, is the number one issue that dampens network capacity in wireless IIoT networks. Devices deployed within the operating range of each other and sending data using the same frequency band vie for the use of that shared radio resource. The higher the device density and the more frequent the data transmission in a network, the more likely multiple messages are sent simultaneously. This leads to packet collisions and eventually, data loss.
Often, self-interference is further complicated by a light-weight, asynchronous medium access control (MAC) protocol such as ALOHA, adopted by many low-power wireless technologies. While reducing complex overheads and, ultimately, power consumption, purely asynchronous systems entail uncoordinated communication among end nodes, which heightens the collision likelihood.
Efficient use of the finite radio spectrum, also known as spectrum efficiency, is key to remedying self-interference and improving overall capacity. This can be achieved by low bandwidth use of a signal and short radio “on-air” time – the time a packet travels from source to destination over the medium. A well-designed radio technology with high spectrum efficiency can offer a huge network capacity of millions of daily messages using only a single base station.
2. Inter-System Interference
Intersystem interference is another major threat to effective operations and the scalability of IIoT communications networks. This problem is often caused by co-existence of multiple radio signals in the same license-free industrial, scientific and medical (ISM) frequency bands.
The 2.4 GHz ISM band is the most widely used frequency in legacy industrial applications and systems and therefore, the most congested one. Wireless technologies using this band are susceptible to significant electromagnetic interference from various sources like Wi-Fi hubs, Bluetooth-enabled cell phones, microwave energy, RF lighting, industrial heaters, and welding equipment. If all of the 2.4 GHz channels in your industrial facility have already been saturated, they aren’t recommended for other IIoT deployments.
Using less crowded sub-GHz ISM bands, for example the 902-928 MHz, can partly mitigate the interference problem. Nevertheless, these bands are now gaining increasing popularity as well. To best tackle near-future exponential traffic in the shared spectrum, system designers should equip their IIoT architecture with a robust, interference resilient wireless technology. Frequency Hopping and channel coding (i.e. Forward Error Correction) are common techniques adopted to improve system robustness.
3. Network setup, management, and maintenance
Managing ten sensors and a thousand sensors are two totally different situations. The theoretical network capacity may sound impressive, but if it involves outrageously complex network configuration and management, manufacturers most likely won’t find the necessary IT competence to scale.
The choice of network topology can decide what will be required from a planning and administration effort. For example, mesh topology, widely adopted by short-range wireless technologies, can be much more sophisticated to manage than the star topology of wide-area solutions. Often times, mesh networks require installing extra devices that act as routers, not to capture important data, but only to achieve the desired coverage. Redundancy is hence an intrinsic part of these networks which can complicate management activities.
Battery life is another critical aspect when it comes to device maintenance and scalability in practice. Maintenance costs account for a substantial proportion of Total Cost-of-Ownership in wireless sensor networks. Imagine how expensive and cumbersome it would be with a thousand field sensors deployed remotely and the need to replace batteries every few days.
4. Communications security
Security comes up in all IIoT conversations, and unless the communications technology is equipped with a versatile, well-proven cryptographic scheme, secure data transfer will be an enormous challenge in ever-expanding IIoT networks.
The Advanced Encryption Standard (AES) is the globally validated standard for message encryption to protect data confidentiality, integrity, and authenticity against eavesdropping and other malicious attempts. Typically, a cipher-based message authentication code (CMAC) can be computed based on AES cipher blocks. Thanks to its low power, high speed, and high throughput, the AES algorithm is highly suitable for wireless sensor networks.
Scalability and interoperability essentially go hand-in-hand, and the key to long-term interoperability is standards. Standards establish consistent and transparent technical models for third-party developers to easily integrate the communication protocol into various IoT devices and application platforms.
An important aspect of a scalable system is its ability to adapt to upcoming technological trends and flexibly evolve with corporate needs. In this context, implementing an industry-standard protocol allows for dynamic future hardware and software changes in the IIoT architecture without incurring downtime or undue delays.
6. Network longevity
Manufacturers in search of a scalable IIoT architecture should also consider the longevity of the communication technology. Longevity reflects stable operations of the IoT system over years without foreseeable disruption due to network shutdowns.
Expensive industrial assets and critical infrastructure often have a very long lifespan (several decades), however certain wireless technologies are already on the verge of phasing out. System designers who choose to ignore the impending sunset of these technologies will probably find themselves having thousands of disconnected devices a few years down the road. Aligning network and device longevity should be a core pillar in a company’s long-term IoT strategy.
Evaluating different communications options
Today, there exists a plethora of wireless communication technologies, each with a different level of scalability and suitability for massive IIoT sensor networks.
High bandwidth technologies like cellular (3G, LTE, etc.) and Industrial WLAN (based on IEEE 802.11 standards) are mainly intended for data-intensive, time-sensitive applications such as security cameras, voice communications, industrial routers, and supervisory control and data acquisition (SCADA) systems. On the downside, the extremely high-power requirement makes them fail to support large-scale, battery-operated sensor networks. Cellular technologies additionally raise considerable concern over network longevity. By 2018, most major global telecoms have announced their 2G shutdowns with the 3G sunset currently rolling out.
Short-range technologies based on IEEE 802.15.4 standards like Zigbee and WirelessHART are more optimized for power consumption due to lower data rates (e.g. 250 kbps). Operating in the 2.4 GHz band, their physical range is limited to few tens of meters due to high rates of signal attenuation and fading. To improve coverage and expandability, these technologies are mostly deployed in a mesh topology. Nevertheless, the relaying mesh functionality can easily shorten battery life, since devices must constantly “listen” for messages that need to be relayed through them.
Scalability of IEEE 802.15.4-based mesh networks can be constrained by several factors. First, though theoretically, nodes can be added as much as needed, inflating complexity in network planning and management – as discussed previously, limits their capacity in practice to a few hundreds of devices. As these networks often cannot scale beyond medium-range applications, their deployments in geographically dispersed industrial campus are hampered. More importantly, IEEE 802.15.4 technologies are exposed to substantial inter-system interference in the congested 2.4 GHz airways that potentially confines network expansion.
Leveraging sub-GHz frequency bands and very low data rates to reduce path loss and improve receiver sensitivity, low power wide area networks (LPWAN) offer excellent range, coverage, and data propagation. Long wavelengths allow LPWAN signals to travel over kilometers and deeply penetrate through walls, buildings, alongside other “heavy” physical obstructions. Therefore, they provide more reliable data transmission than 2.4 GHz signals in structurally dense, rebar environments like mines, refineries, and other manufacturing facilities.
In terms of battery life, LPWAN is the winner thanks to a combination of star topology, deep-sleep mode, and lightweight MAC protocol. Note that both excellent range and power efficiency in LPWAN are achieved at the cost of data rates. These networks are therefore best-suited for condition monitoring, facility management, and predictive maintenance IIoT applications wherein latency is tolerated.
Scalability greatly varies across LPWAN systems due to their difference in network capacity and interference immunity. What’s more, not all LPWAN solutions are based on rigorous industry standards and some of them lack a versatile, valid security scheme like AES. So far, only two camps of LPWAN technologies have been standardized by global Standard Development Organizations and are thus verified for their Quality-of-Service and long-term scalability. One is cellular LPWAN including NB-IoT and LTE-M implementing 3GPP standards. The other is MIOTY™ – an emerging technology implementing the new ETSI standard on Low Throughput Networks – TS 103 357.
As a key takeaway, scalable IIoT architecture requires a communications solution that can support continuous network expansion without compromising system performance. For battery-powered sensor networks in structurally dense, wide area industrial settings, LPWAN provides the most viable scalable solution. With multiple technologies available today, system designers should thoroughly consider critical factors that determine scalability – from both technical and operational viewpoints.