Ethernet offers a higher level of bandwidth and more broadly supported physical layer upgrades for the digital networking already handled by the RS-485, RS-422, and RS-232 connections in fieldbuses such as Modbus® RTU and DeviceNet. In this respect, the transition from fieldbus to Ethernet was more feasible than the previous (and ongoing in some cases) one from point-to-point, 4 mA to 20 mA analog connections to fieldbuses. However, the nondeterminism of Ethernet remained a stumbling block.
By the 1990s, the automation industry began exploring ways around this central obstacle through the development of industrial Ethernet protocols such as Modbus TCP and EtherNet/IP®. These solutions incorporated the Ethernet physical layer as well as a TCP/IP stack, which together could transport data organized in a standard fashion defined by the particular application-level protocol.1
For example, EtherNet/IP uses the reliable common industrial protocol shared by DeviceNet, ControlNet, and CompoNet over standard Ethernet infrastructure to facilitate determinism, in addition to allowing for easy connections from the factory floor all the way up to enterprise IT services.2 Moreover, it can be implemented over the information and control levels in a multitiered automation architecture, as a Rockwell Automation white paper discussed.3
Time synchronization has been a secondary concern on many IT networks until recently. Many other Ethernet-based automation protocols, along with additions to IEEE 802.1—such as audio video bridging—emerged from the years of work that went into making Ethernet suitable for industrial and automotive contexts. Still, there is more work to be done to lay the foundation for the emerging Industrial Internet of Things and continued use throughout the general automation industry.
Back to the Future: How the Age-Old Issue of Timing Will Be Vital to the New IIoT
As a tutorial on deterministic Ethernet presented by Michael Johas Teener of Broadcom pointed out, IT networks have long been designed to carry as much information as possible, with time synchronization only a secondary concern.4 Network specifications lacked mechanisms for keeping track of events on remote devices, unless out-of-band solutions—such as ones predicated on GPS receivers—were used.
To date, the solution has typically been to rely on packet exchange in legacy protocols such as the network time protocol (recently in the news for its exploitation in denial-of-service attacks). This approach has its drawbacks, however. With NTP in particular, a kernel/driver software event as close as possible to the hardware is ideal but not well controlled; only millisecond accuracy is possible.
An event at the Ethernet physical layer is preferable—submicrosecond accuracy could be achieved in this case. The IEEE 1588 precision time protocol was originally proposed in large part to make up for the shortcomings of NTP and simple time protocol, as well as costly OOB management. Optimization for Ethernet TCP/IP networks was also an important specification for PTP.
The National Institute of Standards and Technology has noted the need for superior timing signals to support the interconnection of more devices within the Internet of Things. A February 2015 report from the NIST described the issues with the traditional prioritization of data transfer over tight synchronization.5
“A new economy built on the massive growth of endpoints on the internet will require precise and verifiable timing in ways that current systems do not support,” stated the report abstract. “Applications, computers, and communications systems have been developed with modules and layers that optimize data processing but degrade accurate timing.”
PTP can achieve sub-100 nanosecond timing if the network architecture is fully compliant with IEEE 1588. Compliance entails an Ethernet switch, a PTP slave, and a GPS-enabled PTP grandmaster. Using a GPS satellite, the grandmaster can precisely resolve timestamps and also control a local oscillator that serves as the reference clock for dedicated hardware across the network.6
Incorporation of PTP into industrial Ethernet protocols is already happening:
- EtherNet/IP includes CIP Sync, an extension to CIP that helps EtherNet/IP handle demanding event sequences in motion applications7
- PROFINET® facilitates bandwidth efficiency through a distributed and synchronized clock system that is IEEE 1588-compliant
Other protocols such as EtherCAT® have their own mechanisms for synchronization that achieve similar results as PTP. Going forward, the PTP and its ilk will be critical to ensuring that automation and control systems can provide both the real-time performance and precise timing for mission-critical applications.
1 “10/100M Industrial Ethernet I/O Modules with Modbus.” Acromag, 2005.
2 “Understanding Ethernet Speed and Determinism.” Automation World, 2011.
3 Paul Brooks. “EtherNet/IP: Industrial Protocol White Paper.” Rockwell Automation, 2001.
4 “Deterministic Ethernet.” IEEE, 2012.
5 Marc A. Weiss, John Eidson, Charles Barry, David Broman, Bob Iannucci, Edward A. Lee, Kevin Stanton, Sr, and Leon Goldin. “Time-Aware Applications, Computers, and Communication Systems (TAACCS).” NIST Publications, 2015.
6 Michael Korreng. “IEEE-1588 Precision Time Protocol Synchronization.” Industrial Ethernet Book, Issue 78, 2013.
7 The Future of Industrial Automation. ODVA.