Evolving Towards the Industrial IoT
By Bob McIlvride, Director of Communications, Skkynet
We’ve heard plenty of talk for the past year or two about the Industrial IoT (Internet of Things). The benefits are enticing: leveraging Big Data analysis of production data, connecting seamlessly to remote facilities, shifting capital expenses to operating expenses, and providing vendors, suppliers, and customers with secure access to process data relevant to their needs.
Amidst all the conversation and occasional hype about the Industrial IoT, there has been cautious but steadily increasing implementation. Here and there early adopters are starting pilot and short-term projects. A recent survey report by Machina Research, Lessons Learned from Early Adopters of the IoT: A Global Study of Connected Businesses, suggests that we are witnessing a gradual evolution towards the Industrial IoT.
Focusing on the innovators and early adopters of the IoT, the survey gleaned some useful information which may be helpful for those who have not yet implemented a strategy—and in many cases, those who have. It seems that the majority of early adopters of the IoT took a do-it-yourself approach, and most of them found the IoT more complicated to implement than they expected. “Adopters point to ‘complexity of the IoT solution’ as the largest concern around IoT, a concern that non-adopters have yet to consider fully,” according to the report. Next to complexity were security and difficulty integrating IIoT solutions with existing systems and software.
This experience suggests that an evolutionary, rather than revolutionary, approach to the IIoT makes the most sense. As the technologies improve and mature, as engineers gain experience in implementation, and as management learns how to best leverage these new capabilities, a state of the art will gradually become apparent. In the meantime, here are some questions to consider when evaluating an IIoT solution, to help cultivate an evolutionary approach:
- Is it reasonably simple to implement? The number one complaint from early adopters of Industrial IoT is complexity. Industrial data communications was already complicated on a closed network, even before introducing the idea of connecting to the Internet. Most companies specialize in one area, and offer only part of the solution, counting on someone else to put the pieces together. Those on the cutting edge of the IIoT stress the need to use tools and platforms that are as complete as possible, and that easily interconnect with others.
- Does it support common industrial protocols? OPC Classic and OPC UA play an important role in simplifying and unifying industrial data communications. Any Industrial IoT platform should support OPC, along with common industrial fieldbuses like Modbus, Profibus, HART, DeviceNet, and so on. It should also be able to support more specialized standards like IEC 61850, CAN, ZigBee, and BACnet. It is also a plus if an Industrial IoT platform is compatible with non-industrial standards like ODBC for database connectivity, or DDE for connecting to Excel, if needed, as well as the ability to connect to custom programs.
- Is it secure? Does it require a VPN, which puts all participants onto a local sub-network, giving every one of them access to your whole system? Or is it instead secure by design, limiting users to just data, and only their part of the data set? For Internet connections, can it ensure that all firewall ports stay closed, ensuring zero attack surface? Can it support data tunnelling through proxy servers? Connecting to the IoT is fundamentally different from normal industrial data communication, and requires a correspondingly different architecture.
- How will it impact my existing or legacy systems? Since the introduction of the DCS and PLC in the 1970’s, digital automation has been growing and evolving. While new technologies are constantly being adopted or adapted, many older systems continue to run. Plant management is understandably reluctant to make changes to production systems, having already invested so much engineering, effort, and capital to get them functioning properly. To be accepted in the “If it ain’t broke, don’t fix it” world, an Industrial IoT system should be able to connect to, but not intrude upon, legacy systems. Of course, for new systems, it should do likewise.
- Is it quick enough to be useful? Mission-critical industrial systems are not like consumer or business IoT applications. Performance is crucial. Many IoT systems are built around a relational database, where a response time of a second or two is acceptable, and they send data via HTML or XML, which adds complexity and consumes bandwidth. Although fine for office or home use, these technologies are not sufficient for the Industrial IoT. A better architecture, such as SkkyHub,uses a real-time, in memory database to keep the data moving, a publish/subscribe event-driven model for faster throughput, and a data-centric design with no HTML or XML code for the best performance on the cloud.
- How much time or money will it save me? Traditionally, projects in the industrial sector require large up front capital expenses (CAPEX) and are usually accompanied by long-term commitments. Shifting these costs to operational expenses (OPEX) on the IoT means that you do not need to justify a large capital expenditure over years of returns. Just like a cup of coffee, you buy it, consume it and when you need more, you buy it again. Once in place, the Industrial IoT system can provide further cost savings by supporting Big Data analytics, OEE calculations, and predictive maintenance initiatives.
The implementation of the Industrial IoT is moving more slowly than the hype. But it is actually moving. Rather than a revolution, it seems to be taking shape as an evolution. The experience of early adopters who have walked this path suggests that the benefits are there, particularly for those who watch their step and tread wisely.