Real-time industrial communication protocols are those in which data exchange takes place within a specified amount of time — typically less than 10 ms. But “real-time” systems can be “soft” real-time or “hard” real-time, depending on how rigidly deadlines are enforced. In a soft real-time system, occasional violations of the cycle time, or deadline, for a communication are accepted. On the other hand, in a hard real-time system, there is an absolute limit on response time. If a deadline is missed, a failure or exception will occur. To achieve this, hard real-time systems must be deterministic — that is, the network must guarantee that a message will be transmitted or an event will occur in a specified, bounded amount of time, not faster or slower.
Several industrial Ethernet-based protocols — namely PROFINET IRT (Isochronous Real-Time, also referred to as PROFINET Conformance Class C), SERCOS III, and Ethernet POWERLINK — achieve hard real-time communication by using the principles of isochronous data transmission. In other words, the network relies on coordinated timing via synchronized clocks to reduce cycle time and jitter, providing deterministic behavior that meets the requirements of hard real-time systems.
Segmented and coordinated
In a hard real-time network, the network bandwidth is divided into time slices, with some time slices allocated for hard (isochronous) real-time traffic (IRT), and the rest allocated to standard real-time (RT) traffic. When the isochronous real-time (IRT) time slices come into play, the network buffers and holds the standard real-time (RT) traffic, allowing only the IRT frames to pass through. (The bandwidth, or time slices, reserved for IRT traffic are just long enough for the IRT communications, so that standard RT traffic is not unnecessarily delayed.) Once the IRT frames are clear, the network resumes standard RT traffic.
In PROFINET IRT, this traffic management is made possible by using network switches with special ASICs (rather than standard Ethernet switches). Ethernet POWERLINK also uses a time slot mechanism, together with cyclic polling, to manage isochronous and standard Ethernet traffic, but does so in conjunction with standard Ethernet hubs and switches instead of special hardware. SERCOS III also uses isochronous data transmission, but instead of using switches, each station has an integrated ASIC or FPGA with two communication ports.
Another key to making isochronous real-time communication possible in PROFINET IRT and Ethernet POWERLINK is the use of an extremely accurate, shared clock, per the IEEE 1588 standard. This standard defines a protocol — known as the Precision Time Protocol, or PTP — for “precise synchronization of clocks in measurement and control systems implemented with technologies such as networking communication, local computing, and distributed objects.”
Each switch in the network uses this clock to determine when to activate an IRT time slice and when to allow standard RT traffic. PROFINET IRT goes even further, extending PTP to a protocol referred to as the Precision Transparent Clock Protocol (PTCP), which also computes delays in network switches and cabling. This allows the IRT frames to be sent at an exact, predetermined time. SERCOS III doesn’t use the IEEE 1588 PTP, but instead achieves time synchronization through a master sync message.
EtherCAT and EtherNet/IP also support hard real-time communication. EtherCAT was developed for highly deterministic, real-time communication with extremely short cycle times and synchronization based on IEEE 1588. Ethernet/IP requires the CIP Motion and CIP Sync extensions to achieve deterministic real-time behavior.
Isochronous real-time communication is typically characterized by cycle times of less than 1 ms and jitter less than 1 μs. These requirements are found in motion control applications that involve coordinated movements (especially multi-axis motion control) or that require precisely timed, repeatable communications or actions that need to be carried out simultaneously.