Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC TR 62251: Quality of Service in Industrial Communication Networks
A technical framework for evaluating and managing QoS parameters in industrial automation communication systems
IEC TR 62251, published in 2003 as a Technical Report by IEC Technical Committee 65 (Industrial-Process Measurement, Control and Automation), addresses the critical topic of quality of service (QoS) in industrial communication networks. As manufacturing systems transition from centralized control architectures to distributed, networked automation, the performance characteristics of the underlying communication infrastructure become decisive factors in system reliability, safety, and productivity. This report provides a structured methodology for defining, measuring, and managing QoS parameters in industrial environments where real-time determinism and high availability are essential requirements.
Unlike office IT networks where best-effort delivery is acceptable, industrial communication networks must guarantee deterministic behavior. A missed deadline in a motion control loop or a safety-critical message can cause product defects, equipment damage, or personnel injury. IEC TR 62251 provides the framework for specifying these guarantees.
QoS Parameters and Classification Framework
IEC TR 62251 defines a comprehensive set of QoS parameters organized into three principal categories: temporal performance, reliability, and availability. Temporal parameters include transmission delay (the time from when a message is submitted to the communication layer at the source to when it is delivered at the destination), delay jitter (the variation in transmission delay between successive messages), and throughput (the amount of data successfully transferred per unit time). For industrial applications, these parameters must be specified as worst-case bounds rather than averages, since the control system must function correctly under all operating conditions including peak traffic loads.
Reliability parameters in the report include message loss ratio (the proportion of transmitted messages not successfully delivered within the required time bound), residual error rate (undetected bit errors despite protocol error detection mechanisms), and connection establishment failure probability. Availability parameters describe the proportion of time the communication service meets its specified QoS targets, expressed as a percentage (e.g., 99.999% availability for safety-critical communication links).
The report classifies industrial applications into QoS classes corresponding to the stringency of their communication requirements. Class A applications (motion control, safety systems) require isochronous real-time communication with maximum delays below 1 ms and jitter below 1 microsecond. Class B applications (closed-loop process control) require hard real-time performance with delays of 1-10 ms. Class C applications (SCADA, configuration, monitoring) can tolerate soft real-time or best-effort performance with delays of 10-100 ms or more.
| QoS Class | Application Examples | Max Delay | Max Jitter | Availability |
|---|---|---|---|---|
| A – Isochronous | Motion control, drives, safety | < 1 ms | < 1 micro-s | 99.999% |
| B – Hard real-time | Process control, robotics | 1 – 10 ms | < 100 micro-s | 99.99% |
| C – Soft real-time | SCADA, batch control | 10 – 100 ms | < 10 ms | 99.9% |
| D – Best effort | Monitoring, logging, configuration | 100 ms – 1 s | Not critical | 99% |
A common mistake in industrial network design is specifying average delay targets rather than worst-case bounds. In a Class A motion control application, even a single delayed cycle (exceeding the 1 ms bound) can cause a servo drive to fault, shutting down an entire production line. The QoS specification must use maximum (not mean) values for critical parameters.
QoS Mechanisms and Protocol Considerations
IEC TR 62251 reviews the QoS mechanisms available in common industrial communication technologies. For industrial Ethernet (PROFINET, EtherCAT, EtherNet/IP, SERCOS III), the report addresses the role of IEEE 802.1Q VLAN tagging for traffic prioritization, the use of time-sensitive networking (TSN) mechanisms including IEEE 802.1AS time synchronization and IEEE 802.1Qbv scheduled traffic, and the configuration of communication cycle times to match application requirements. The report emphasizes that QoS configuration must be consistent across all network devices including switches, routers, and end devices to prevent QoS configuration gaps that can cause unpredictable behavior.
For fieldbus systems (PROFIBUS, Foundation Fieldbus, CAN-based protocols), the deterministic behavior is typically built into the medium access control method. Token-passing and master-slave protocols inherently guarantee bounded delay through controlled access. The report provides guidance on calculating worst-case communication cycle times based on the number of nodes, message sizes, and baud rate, enabling engineers to verify that the network topology supports the required QoS class before deployment.
Wireless industrial communication receives dedicated attention, including wireless LAN (IEC 62591, WirelessHART) and cellular-based solutions. The report notes that wireless links introduce additional QoS challenges: variable link quality due to fading and interference, potential for hidden-node collisions, and the overhead of retransmission and channel hopping. QoS mechanisms for wireless include adaptive data rate selection, frequency diversity through channel hopping, and redundant access points for coverage reliability.
| Technology | MAC Method | Min Cycle Time | Inherent QoS | TSN Ready |
|---|---|---|---|---|
| PROFINET IRT | Scheduled (TDMA) | 31.25 micro-s | Class A | Yes |
| EtherCAT | Summation frame | 12.5 micro-s | Class A | Under development |
| EtherNet/IP (CIP Sync) | CSMA/CD + TSN | 100 micro-s | Class B | Yes |
| SERCOS III | Time-division | 31.25 micro-s | Class A | Yes |
| PROFIBUS DP | Token-passing | 1 ms | Class B | N/A |
| WirelessHART | TDMA + FHSS | 10 ms | Class C | N/A |
Time-Sensitive Networking (TSN) is revolutionizing industrial communication by enabling standard Ethernet hardware to deliver Class A QoS. The IEEE 802.1 TSN task group standards, referenced by IEC TR 62251, provide deterministic scheduling, precise time synchronization, and seamless redundancy that were previously achievable only with specialized fieldbus technologies. TSN converged networks can carry real-time control traffic alongside IT traffic on a single infrastructure, reducing cabling and hardware costs.
Engineering Design Insights for Industrial Communication QoS
Network topology selection is among the most consequential decisions in industrial communication design. Line topologies are simple and cost-effective but accumulate delay across each hop, potentially exceeding QoS bounds for time-critical traffic at the end of the chain. Ring topologies with redundancy protocols (MRP, DLR) provide fault tolerance at the cost of increased worst-case delay during reconfiguration. Star topologies minimize deterministic delay variation but require more cabling and switch ports. For Class A applications, daisy-chaining more than 5-10 devices on a single line segment is generally not recommended.
Traffic engineering is equally critical. Industrial networks carry multiple traffic types with different QoS requirements: cyclic real-time data (I/O, drives), acyclic real-time data (alarms, parameter changes), non-real-time data (configuration, diagnostics), and best-effort traffic (logging, video). Each type must be classified, marked, and queued appropriately at every network node. The report recommends allocating dedicated priority queues for real-time traffic with strict priority scheduling, while using weighted fair queuing for non-real-time traffic to prevent starvation.
Network dimensioning must account for both normal operation and fault scenarios. Under normal conditions, link utilization should not exceed 30-40% for real-time traffic to accommodate transient bursts and fault recovery traffic. Redundancy mechanisms must ensure that a single link or device failure does not degrade QoS below the requirements of any active application. The report recommends using the R-A-T framework (Redundancy, Availability, and Testability) as a structured approach to designing resilient industrial communication systems.
Finally, the report emphasizes that QoS verification through systematic commissioning tests is essential. The test plan should include latency measurements under normal load, latency measurements under peak load (simulated using traffic generators), failover timing during link or device failure, and long-term availability monitoring (minimum 72 hours continuous). Only through such testing can engineers confirm that the installed system meets the QoS requirements specified during the design phase.
Q1: What are the most critical QoS parameters for industrial communication?
A: The three most critical parameters are maximum transmission delay (worst-case latency), delay jitter (latency variation), and availability percentage. For motion control applications, delay and jitter bounds below 1 ms and 1 microsecond are typical. For process control, delays of 1-10 ms are acceptable with moderate jitter. Availability requirements range from 99.9% (standard process) to 99.999% (safety-critical applications). The message loss ratio and residual error rate are also important for applications with high integrity requirements.
A: The three most critical parameters are maximum transmission delay (worst-case latency), delay jitter (latency variation), and availability percentage. For motion control applications, delay and jitter bounds below 1 ms and 1 microsecond are typical. For process control, delays of 1-10 ms are acceptable with moderate jitter. Availability requirements range from 99.9% (standard process) to 99.999% (safety-critical applications). The message loss ratio and residual error rate are also important for applications with high integrity requirements.
Q2: Can standard office-grade Ethernet switches be used in industrial communication networks?
A: Generally no, for three reasons. First, office switches do not support the deterministic TSN scheduling mechanisms required for Class A/B QoS. Second, office switches lack the environmental ruggedness (extended temperature, vibration, EMI immunity) expected in industrial environments. Third, office switches typically have limited QoS queue depth and configuration options. Industrial-grade managed switches with TSN support, extended temperature rating, and redundant power supplies are the appropriate choice for production networks.
A: Generally no, for three reasons. First, office switches do not support the deterministic TSN scheduling mechanisms required for Class A/B QoS. Second, office switches lack the environmental ruggedness (extended temperature, vibration, EMI immunity) expected in industrial environments. Third, office switches typically have limited QoS queue depth and configuration options. Industrial-grade managed switches with TSN support, extended temperature rating, and redundant power supplies are the appropriate choice for production networks.
Q3: How does wireless communication affect QoS compared to wired industrial networks?
A: Wireless introduces additional QoS challenges: higher and variable latency (typically 5-50 ms for industrial wireless), higher jitter due to retransmissions and channel hopping, susceptibility to interference from other wireless devices, and lower effective throughput (typically 50-70% of nominal due to protocol overhead). WirelessHART and ISA100.11a address these through TDMA scheduling, frequency hopping, and mesh networking. For Class C applications, wireless can be an effective solution, but Class A applications still require wired connections in most cases.
A: Wireless introduces additional QoS challenges: higher and variable latency (typically 5-50 ms for industrial wireless), higher jitter due to retransmissions and channel hopping, susceptibility to interference from other wireless devices, and lower effective throughput (typically 50-70% of nominal due to protocol overhead). WirelessHART and ISA100.11a address these through TDMA scheduling, frequency hopping, and mesh networking. For Class C applications, wireless can be an effective solution, but Class A applications still require wired connections in most cases.
Q4: What is the recommended approach to migrating existing fieldbus networks to industrial Ethernet with QoS guarantees?
A: A phased approach is recommended: (1) audit existing applications and classify their QoS requirements using the Class A-D framework, (2) assess the existing cabling infrastructure for suitability (most PROFIBUS and Foundation Fieldbus cabling will need replacement), (3) select a TSN-capable industrial Ethernet technology that matches the highest QoS class required, (4) conduct a pilot installation on a non-critical production line, and (5) validate QoS compliance through systematic measurement before full deployment. The migration of individual devices can be phased if proxy/gateway devices are used to bridge between legacy fieldbus segments and the new Ethernet backbone.
A: A phased approach is recommended: (1) audit existing applications and classify their QoS requirements using the Class A-D framework, (2) assess the existing cabling infrastructure for suitability (most PROFIBUS and Foundation Fieldbus cabling will need replacement), (3) select a TSN-capable industrial Ethernet technology that matches the highest QoS class required, (4) conduct a pilot installation on a non-critical production line, and (5) validate QoS compliance through systematic measurement before full deployment. The migration of individual devices can be phased if proxy/gateway devices are used to bridge between legacy fieldbus segments and the new Ethernet backbone.
