This is the second article by Derek Long, Head of Telecoms and Mobile, Cambridge Consultants. In his first post, Long argued that mobile network operators face challenges in addressing the industrial IoT market.
In this second article, Long assesses the technical options for achieving the required radio network coverage, security and performance for industrial IoT applications.
Connecting the Industrial IoT
The Internet of Things has been spoken of for many years and is usually spoken of in terms of other fashionable tech trends, including smart homes, smart cities and connected cars. But perhaps one of the most important applications of IoT is Industrial IoT (IIoT), which at its purest level means the integration of operations technology in the factory with IT in the office. Or to put it another way: connected manufacturing.
But Industrial IoT is much easier said than done, particularly the radio connectivity elements. In this piece I’ll detail the radio-based technologies for providing that connectivity in the demanding environments of a modern manufacturing facility.
To take a step back for a moment, as an enterprise or as a provider of services to enterprises, it’s important to consider many factors before introducing the higher levels of efficiency and automation which are enabled by the Industrial IoT. Questions such as: what to do with all the data generated by sensors being introduced to modern machinery, what level of automation and control to give the overlying service, whether or not to automate performance improvements with machine learning and how to handle practical questions such as provisioning, management, remotely maintenance, device upgrades and more.
Fundamental to all of these is how to ensure reliable connectivity between the devices in the manufacturing facility and the control service, in the highly demanding industrial environment, which may experience significant RF interference from a number of uncontrolled sources, and the demanding RF propagation environment found in the large metallic structures which form many modern manufacturing facilities.
When considering which communication technology to use as the basis for an Industrial IoT implementation, there are a number of factors to consider. These include:
– Licensed or unlicensed spectrum?
– Appropriate robustness requirements
– Security
– Throughput and other practical requirements
The answers to these questions very quickly reduces the choice of appropriate connectivity technologies. A quick analysis of potential connectivity technologies shows that there are around 20 to choose from at present, ranging from standardised, mass-market technologies, to dedicated, proprietary solutions.
Licensed or unlicensed spectrum?
Using a technology that requires licensed spectrum quickly leads the enterprise into the hands of mobile network operators that own spectrum. Furthermore, these companies predominantly use 3GPP-based technologies that require a SIM for operation. The SIM provides security, as well as being the basis for provisioning and other services. Mobile operators may also have beneficial business models, such as a subscription model, which in the short term removes the need for investment in connectivity infrastructure but in the long term creates a dependency on the operator.
Appropriate robustness requirements
Secondly, robustness is a critical issue. In the case of licensed spectrum, it’s necessary for the operator or an integrator to perform some form of cell and site planning, in addition to transmission planning. The purpose of such planning is to limit the interference the system creates for itself and to limit interference from other systems. If the choice is for unlicensed spectrum (which generally means the ISM bands at 868, 916, 2.4GHz and 5GHz) then the interference environment is out of the enterprise’s control, and it becomes necessary to use a technology that can withstand interferers. Interferers might be other communications systems, such as WiFi, but could also include non-communications related sources such as radar, microwave ovens and heavy machinery.
Robustness on the air interface can be included in the system, through error correction coding, end-to-end service assurance mechanisms and also through time and frequency agility, allowing the system to adapt itself continuously to the ambient interference environment.
Security
Security is such a vital topic that it justifies an article in its own right. There is however one important point to consider, which is that security must be considered at various levels in the system. For example, the sensor/actuator in the device must be secure, the connectivity link between device and IT system must also be secure, and finally, the end-to-end system including the overall service must be secure.
If there is potential for incorrect sensor data to be forwarded to the control service, then the potential for creating havoc is huge.
This level of security is necessary even if the enterprise is only considering a network of sensors. Many deductions about the health of a company can be made from sensor data generated on the factory floor. If there is potential for incorrect sensor data to be forwarded to the control service, then the potential for creating havoc is huge. Of course having an agile connectivity network, as described above, adds to the overall system security as it increases the complexity of the scanners or blockers required.
Practical requirements
The final set of considerations relate to overall system practicalities. Whilst it might be interesting to select the latest low-latency, high-reliability technology, this may not always be necessary. A carefully designed system will balance the processing requirements in the device with those in the service. There are a number of points to be considered such as the throughput required, latency requirements (the time between a message from the device and the response from the service) and whether the system will be used for the distribution of software upgrades.
The Industrial IoT is likely to change the world of manufacturing significantly, from merely improving the level of automation and efficiency in the factory, all the way to providing fully automated start-to-finish product lifecycle support and management, which in turn could transform the way we buy, use and own products in future.
At a fundamental level, IIoT systems are the integration between the operations technology of the production line and the information technology of the business process system. Despite their different heritages, these two technology domains are soon to become one.
