Having just put this month’s cover story on the imminent commercialization of Ethernet-APL to bed, I’m left wondering if the process industries are really prepared for an advanced physical layer (APL) that will bring Ethernet that elusive “last mile” to two-wired, loop-powered field instruments in its many potentially explosive environments.
On the one hand, APL is a significant technical achievement that dramatically increases the bandwidth available for digital instrument communications compared with its predecessors. Also, since it’s Ethernet, APL simplifies system architecture from the field device all the way to enterprise business systems and the cloud without the need for gateways or other protocol conversion devices. And, while this direct connection from cloud to sensor is in some ways a frightening prospect, it also opens a way for instrument and system manufacturers to finally implement long-neglected, top-to-bottom cybersecurity measures that extend all the way to digital field devices.
Compelling value propositions all, but it seems we’ve been here before. The primary all-digital alternatives, FOUNDATION fieldbus and Profibus PA, are based on technology now more than 20 years old—and neither one successfully displaced the 4-20mA analog + HART hybrid option that's 15 years older still.
Over the past 10 years, even as APL Project developers soldiered on, control systems suppliers focused on the configurable input/output (I/O) solutions that successfully decoupled system hardware and software design tasks. This accomplished the central goals of streamlining system project engineering, and removed control engineering from the critical path of project execution. All the major distributed control systems providers—and the engineering services firms that execute major projects—have now reoriented their engineering practices around this approach.
A perhaps unintended consequence of the move to configurable I/O is that it also solidified the continued centrality of analog 4-20mA + HART communications. Sure, DCS vendors and EPCs can pivot once again, filling those off-the-shelf junction boxes with APL field switches instead of configurable I/O, but how long will that transition take? And perhaps more importantly, will end users demand it?
“Initially, we expect customers to continue to make decisions based on comfort, cost and capability,” notes Emerson’s Andy Kravitz in this month’s cover story, “and their selection of devices will likely not change due to the release of a new technology such as Ethernet-APL.” This is likely to change over time, as instrumentation and system vendors ramp up support of APL, Kravitz continues, and “customers will start to select Ethernet-APL devices due to the potential for increased capability, security and flexibility.”
Better prepared for complexity?
More than one pundit has attributed the slow uptake of first-generation fieldbus options to their increased complexity. Techs were comfortable with their handheld multimeters and HART communicators, but the jump to all digital proved a bridge too far. Certainly, the convergence of IT and OT networks continued to accelerate over the past 20 years, but are the people who make this stuff work in the field any better prepared today for a digital field than they were then?
On the other hand, I also seemed to remember early fieldbus proponents focusing on wiring savings as a key value proposition—that daisy-chaining multiple instruments on a single cable would reduce wiring costs. Even though that proved a false economy, as most opted for star topologies to limit the risk of multiple instrument failures if one segment failed, it’s certainly a thin argument compared with that for APL.
Only time will tell.