Private wireless networks are playing a big and growing role in the realization of Industry 4.0. Many advanced industrial use cases require ultra-reliable, low latency wireless connectivity to ensure real-time data exchanges between machines, IoT devices and AI/ML analytics to power automation. Based on 4.9G/LTE and 5G, these wireless networks employ technologies that make it possible to meet the most stringent performance requirements.
Despite being standards-based, however, private wireless networks based on 4.9G/LTE or 5G are not all created equal. The standards ensure that any device that is based on the radio technology can communicate with the network, but it doesn’t determine the level of performance that is delivered. Conforming to the standard is only the beginning. How the components and solution architecture are executed can make for very unequal results.
High availability
In industrial OT environments, many of the use cases require up to five 9s (99.999%) uptime. Machine-control communications, for instance, are critical for reliable and predictable operations in assembly line environments. Even brief periods of downtime can have significant cost implications, especially when viewed on an annual basis. Unreliable network availability can also cause safety issues, especially with autonomous mobility applications such as an autonomous forklift, crane, straddle carrier or 320-ton ore hauler.
Compared with industrial operational technology (OT) business- and mission-critical use cases, enterprise day-to-day IT use case requirements have more modest requirements for how available the network needs to be. Measured in percentage terms, it is generally considered sufficient for enterprise networks, such as Wi-Fi, to provide 99 to 99.95% availability, which is why OT use cases require the use of different wireless connectivity technologies. Private wireless networks can easily meet 99.99% availability, and when required, can even go beyond.
Typical features of industrial-strength private wireless networks solutions for OT applications include multiple layers of redundancy. But for more stringent performance levels, suppliers can beef up the solution capabilities and the customer network design by using multiple radio spectrum bands to create multi-layer radio connectivity, geo-spatial separation of redundant small cells, and a third level of redundancy from, for example, a public network core.
Looking to the future, using 5G SA, private wireless networks will be able to evolve to six and seven 9s availability, which may be needed to support future critical machine communications use cases.
Experience is essential
Often the missing ingredient in achieving very high availability is experience. Many private wireless solution providers have little or no experience in delivering end-to-end solutions for these critical enterprise networks. Specialized high availability wireless networks have existed for decades, such as GSM-R wireless networks used by the railway industry or public safety needs. The experience gained servicing these networks favors a much smaller group of network suppliers that have proven experience delivering these kinds of mission-critical services.
Designing private wireless networks for high availability means integrating the network with the industrial applications that rely on it. This is especially critical where edge clouds are being used for edge computing applications – often needed for meeting high reliability and low latency. The network supplier must have extensive experience integrating specific industrial partner elements and software. This is systems integration work that requires the supplier to have solution services capabilities, including pre-integrated and tested solutions that can tackle the key use cases of each segment with a range of preferred partners’ assets.
Ultra-reliability
Although connected, the issue of performance reliability (aka predictability or determinism) is distinct from availability. Beyond the issue of network uptime, reliability addresses the issue of consistent service performance. Radio design plays a major role in delivering reliable private wireless, but that is not all.
Unfortunately, many private wireless vendors use Femtocell (or inexpensive small cells) ‘all-in-one’ chipsets available from chipset vendors for their radios. These vendors not only provide the reference design but also the baseline software stack. This means virtually any company, not even a radio expert, can work with a contract manufacturer to rapidly build a range of private wireless radios based on these chipsets and baseline software.
The use case for Femtocells is not generalizable to many Industry 4.0 applications. Not only were they designed for email and web browsing in homes with very few users, but it was always assumed that the macro cellular coverage could fill in when the in-home Femto cell failed. The macro radio cell from mobile operators uses a layered approach that covers multiple spectrum bands including older standards such as 2G, GRPS and 3G, so there are multiple redundancies at the macro level, hence Femtocells use in such networks is ok and serves the purpose.
Private wireless networks do not have this wide range of spectrum band options nor the public network to fall back on. In addition, the radios are often deployed in very challenging radio environments such as underground mines and factories and warehouses with high ceilings. And the number of potential IoT sensors and other devices is much larger.
