Fact-checked by Grok 2 weeks ago

High-availability Seamless Redundancy

High-availability Seamless Redundancy (HSR) is a redundancy protocol for Ethernet-based networks that provides seamless failover with zero recovery time in the event of a single network component failure, ensuring continuous data transmission without interruption. Defined in Clause 5 of the international standard IEC 62439-3, HSR achieves this by duplicating each transmitted frame at the source node and sending the copies simultaneously over two counter-rotating paths in a ring or meshed topology, with the destination node accepting the first arriving frame and discarding subsequent duplicates to prevent loss or delay.^[1]^[2] Developed for high-availability automation networks, HSR supports real-time applications requiring deterministic latency and short cycle times, such as substation automation in smart grids, industrial control systems, and railway signaling. It operates with doubly attached nodes (DANs) that integrate two Ethernet ports for the redundant paths, while singly attached nodes can connect via a dedicated HSR interface (DANH) to maintain compatibility. The protocol uses supervision frames to monitor network health and detect failures, inherently avoiding loops without relying on spanning tree algorithms like RSTP, which can introduce recovery delays.^[3]^[2]^[1] Introduced in the 2010 edition of IEC 62439-3 as an alternative to the Parallel Redundancy Protocol (PRP), HSR offers a cost-effective redundancy solution for ring topologies by eliminating the need for fully parallel network infrastructures, though it doubles the traffic load on the network. Subsequent updates, including the 2021 edition, have enhanced its support for meshed configurations, broadening its applicability in mission-critical environments.^[4]^[1]^[5]

Introduction

Definition and Objectives

High-availability Seamless Redundancy (HSR) is a standardized protocol for Ethernet networks that delivers seamless redundancy by duplicating data frames and transmitting them along multiple paths, ensuring zero recovery time in the event of a single network component failure. Defined in Clause 5 of IEC 62439-3, HSR operates as a self-managed, application-protocol-independent mechanism that eliminates single points of failure without requiring a central supervisor, making it suitable for ring-based topologies where continuous operation is paramount.^[6]^[7] The primary objectives of HSR include maintaining uninterrupted data flow with recovery times on the order of microseconds, thereby supporting real-time communication in mission-critical environments. It achieves this through a duplicate discard process that prevents network loops by identifying and removing redundant frames at the receiving end, based on unique identifiers, while preserving the integrity of the original traffic. This design prioritizes high availability and fault tolerance, enabling deterministic performance without packet loss during failures.^[7]^[6] HSR finds key applications in substation automation systems compliant with IEC 61850, where it ensures reliable, high-speed communication for protection relays and process-level data exchange. It is also employed in synchronized drive systems, such as those in printing machines, high-power inverters for industrial control, and smart grid networks requiring robust redundancy for time-sensitive operations.^[8]^[6]^[2] In contrast to traditional spanning tree protocols like RSTP, which involve convergence delays of milliseconds to seconds to reconfigure paths and avoid loops, HSR provides instantaneous failover through parallel transmission, offering superior performance for applications intolerant to even brief interruptions.^[9]^[2]

Historical Development and Standards

The development of High-availability Seamless Redundancy (HSR) originated from efforts to enhance fault tolerance in industrial Ethernet networks, particularly for substation automation. Initially proposed as HASAR, an acronym derived from the initials of the key contributing companies (Hirschmann, ABB, Siemens, Alstom, and RuggedCom), around 2008, the protocol was discussed in IEC TC57 WG10 meetings as a ring-based redundancy solution for high-availability applications. Key contributors included Hirschmann, ABB, Siemens, Alstom, and RuggedCom, who collaborated on its foundational concepts during 2008-2010 to address zero-recovery-time needs in harsh environments.^[10]^[11] By 2010, the protocol was renamed HSR and formalized as Clause 5 in the first edition of IEC 62439-3, which specified it alongside the Parallel Redundancy Protocol (PRP) for seamless failover in Ethernet-based automation networks.^[4] This edition, published in February 2010, introduced HSR's core principles of frame duplication and discard to prevent loops without interrupting traffic. Subsequent updates refined the standard: the third edition in 2016 enhanced interoperability and performance specifications, while the fourth edition (Edition 4.0) in 2021—followed by a corrigendum in 2023—incorporated improvements such as enhanced PTP integration and multi-vendor compatibility, building on existing support for meshed topologies.^[1]^[12]^[13] In 2016, HSR was integrated into IEC/IEEE 61850-9-3, a precision time protocol (PTP) profile for power utility automation, enabling sub-microsecond synchronization over redundant HSR rings.^[14] Key milestones include a 2012 IEEE paper detailing HSR's duplicate discard mechanism, which ensures loop-free operation by identifying and removing redundant frames based on node-specific tags.^[15] Additionally, hybrid modes combining HSR with PRP emerged to support mixed ring and parallel topologies, allowing seamless connectivity between HSR rings and PRP networks via RedBox devices.^[16] Post-2023, no major revisions to the core HSR standard occurred, but advancements focused on practical implementation and interoperability. For instance, the Flexibilis FES-HSR switch underwent successful multi-vendor testing in 2012 and later events, demonstrating gigabit-speed compatibility in HSR rings.^[17] In software, a 2024 Linux Plumbers Conference presentation highlighted the current status and ongoing development of the Linux HSR driver, emphasizing enhanced support for zero-switchover testing via QEMU emulation.^[18] In 2025, advancements continued with vendor-specific enhancements, such as Cisco's introduction of HSR-PRP hybrid mode in IOS XE 17.14.1, supporting seamless integration in mixed redundancy environments.^[16]

System Architecture

Topology Configurations

High-availability Seamless Redundancy (HSR) primarily employs a ring topology, where HSR-capable devices, known as doubly attached nodes for HSR (DANHs), connect in a closed loop using two Ethernet ports per device.^[19] Each DANH transmits duplicate frames simultaneously in both clockwise and counterclockwise directions across the ring, ensuring seamless redundancy without interruption upon link or node failure.^[20] This configuration, defined in IEC 62439-3 Clause 5, supports full-duplex Ethernet links and is optimized for industrial applications requiring zero recovery time.^[21] To extend beyond a single ring, HSR supports mesh networks formed by interconnecting multiple rings through redundancy coupling devices, such as pairs of QuadBoxes, which prevent single points of failure while maintaining HSR protocol integrity.^[22] Linear extensions can be achieved by attaching non-ring segments via RedBoxes, allowing integration of linear topologies into the primary ring without disrupting redundancy.^[19] Hybrid configurations with Parallel Redundancy Protocol (PRP) are also possible, as in HSR-PRP mode, which bridges an HSR ring to a PRP mesh network using dual RedBoxes connected to PRP LAN A and LAN B, introduced in Cisco IOS XE 17.14.1 for enhanced flexibility in mixed environments.^[23] RedBoxes play a crucial role in HSR topologies by enabling the attachment of single-port legacy equipment, termed singly attached nodes (SANs), to the ring. These devices act as proxies, converting SANs into virtual DANs (VDANs) by handling duplicate frame transmission and reception, thus allowing non-HSR-compliant devices to benefit from the ring's redundancy without requiring hardware upgrades.^[20] QuadBoxes, an extension of RedBoxes, facilitate coupling between rings or hybrid setups by providing four ports for redundant connections.^[21] HSR topologies require all core nodes to be doubly attached via dedicated ring ports (Port A and Port B), ensuring continuous duplicate paths.^[19] Implementations typically support up to 50 nodes in a ring, though IEC 62439-3 Annex C recommends a maximum of 16 for optimal performance to minimize timing inaccuracies and maintain low latency.^[24]

Node and Device Roles

In High-availability Seamless Redundancy (HSR) networks, the fundamental building block is the Doubly Attached Node for HSR (DANH), which features two independent Ethernet ports designated as Port A and Port B. These ports facilitate the duplication and transmission of frames in opposite directions along counter-rotating paths, ensuring redundancy without requiring external switches for basic ring operation.^[2]^[20] DANHs are designed to originate and terminate HSR traffic, meaning they generate duplicate frames upon transmission and discard superfluous duplicates upon reception to prevent network flooding. In contrast, Singly Attached Nodes (SANs), which lack native HSR capabilities, connect to the HSR ring indirectly through specialized devices called RedBoxes. RedBoxes act as intermediaries, integrating non-HSR equipment into the redundant topology while maintaining the protocol's seamless failover properties.^[19]^[25] Internally, DANHs and RedBoxes incorporate cut-through switching mechanisms to forward frames with minimal latency, often achieving forwarding times under 3 microseconds by beginning transmission before the entire frame is received. RedBoxes typically include at least three ports: two HSR redundancy ports (A and B) for ring attachment and an interlink (IL) port to interface with non-HSR network segments, enabling the extension of redundancy to legacy Ethernet devices.^[26]^[27]^[2] To support hybrid topologies, HSR nodes and RedBoxes can interconnect with Parallel Redundancy Protocol (PRP) networks, often via dual RedBox configurations where the two HSR ports link to the ring and the IL port connects to one of the PRP LANs (A or B), allowing seamless integration across redundancy schemes defined in IEC 62439-3.^[19]^[28]

Operational Principles

Frame Transmission and Reception

In High-availability Seamless Redundancy (HSR), frame transmission begins at the source node, known as a Doubly Attached Node for HSR (DANH), which duplicates each unicast or multicast frame and sends identical copies simultaneously over its two dedicated ring ports, labeled Port A and Port B.^[19] This dual transmission creates counter-rotating paths around the ring topology: frames on Port A travel counterclockwise, while those on Port B travel clockwise, ensuring that if one path fails due to a link or node fault, the alternate path delivers the frame without interruption.^[29] Each transmitted frame includes an HSR tag appended to the Ethernet frame, which contains a sequence number incremented for each new frame from the source to facilitate later duplicate detection.^[21] This process adheres to the requirements of IEC 62439-3 Clause 5, enabling zero-recovery-time redundancy in fault-tolerant Ethernet networks.^[6] At the destination DANH, frame reception involves monitoring both incoming ports for the duplicated frames, which typically arrive in quick succession during normal operation. The receiving node processes only the first arriving frame by stripping the HSR tag and forwarding the original payload to the upper-layer application, while subsequent identical frames are discarded based on matching the source MAC address and the sequence number in the HSR tag.^[19] This selective acceptance ensures that the network delivers each unique frame exactly once, preventing data duplication at the application level and maintaining seamless operation even if one frame is delayed or lost due to a transient fault.^[21] Intermediate DANHs forward received frames on their outgoing port without processing the payload, propagating the duplicates around the ring until they reach the intended destination or return to the source.^[6] HSR's design inherently prevents network loops through the automatic discard of duplicates, eliminating the need for spanning tree protocols like RSTP. For unicast frames, the destination discards the second copy, and any frame circulating back to the source is also dropped after completing the ring traversal.^[29] Broadcast and multicast frames are handled similarly, with each DANH forwarding the first received copy on its alternate port while discarding subsequent duplicates to avoid broadcast storms, though this increases bandwidth usage by up to 200% compared to standard Ethernet due to the doubled traffic.^[19] This loop-free mechanism relies on the consistent duplicate filtering at every DANH, as specified in IEC 62439-3, ensuring deterministic behavior in ring topologies without additional configuration.^[6] Special cases in HSR frame handling include support for VLAN-tagged frames, allowing prioritization via the 802.1Q priority code point (PCP) subfield.^[6] For integrating non-HSR devices, such as Singly Attached Nodes (SANs), RedBoxes act as proxy nodes that connect the SAN to the HSR ring; they duplicate outgoing frames from the SAN onto both ring directions and filter incoming duplicates before forwarding a single copy to the SAN, effectively virtualizing it as a Doubly Attached Node (VDAN).^[29] This proxy functionality, also defined in IEC 62439-3 Clause 5, enables legacy Ethernet devices to participate in the redundant network without modification.^[19]

Redundancy Mechanisms and Modes

High-availability Seamless Redundancy (HSR) achieves fault tolerance through a core mechanism of duplicating each data frame and transmitting it simultaneously along two counter-rotating paths in a ring topology, ensuring that the receiving node accepts the first arriving frame while discarding any subsequent duplicate to prevent loops.^[30] This parallel path approach provides zero recovery time upon failure of a single link or node, as the alternate path delivers the frame without interruption or reconfiguration.^[31] HSR operates in several modes to optimize traffic and compatibility. The standard mode H (High availability) forwards all frames on both ports unless a duplicate is detected, suitable for full ring redundancy but resulting in doubled traffic.^[32] Mode U (Unidirectional) reduces traffic in linear or tree-like setups by forwarding unicast frames only toward their unique destination and supervision frames in one direction, minimizing unnecessary duplication while maintaining redundancy supervision. Mode X extends this for proxy nodes (RedBoxes) interfacing non-HSR devices, where multicast and broadcast frames are not forwarded if a duplicate has already been received, further optimizing mixed-traffic environments.^[33] Failover in HSR is triggered by the detection of link or node failures via periodic supervision frames, which are periodically multicast at configurable intervals (typically milliseconds to seconds, depending on implementation) to monitor path integrity; absence of expected supervision frames from a neighbor indicates a fault, prompting immediate reliance on the surviving path.^[34]^[21] Certain protocol frames, such as Precision Time Protocol (PTP) and Link Layer Discovery Protocol (LLDP), are handled without duplication: PTP frames are modified by each node for time correction, making them non-duplicates, while LLDP frames are typically sent singly to avoid redundant neighbor announcements.^[35] The duplicate discard process, relying on sequence numbers in the HSR tag, ensures only the initial frame is processed at the destination. For hybrid topologies, HSR integrates with Parallel Redundancy Protocol (PRP) in HSR-PRP mode, allowing rings to connect to star networks via proxy devices that translate between protocols, as introduced in Cisco IOS XE 17.14.1 for enhanced flexibility in industrial setups.^[19]

Frame Structure

HSR Tag Composition

The HSR tag is a 6-octet structure inserted immediately after the source MAC address (and after any VLAN tag if present) in the Ethernet frame header, preceding the original payload. Defined in IEC 62439-3 Clause 5, this tag facilitates seamless redundancy by enabling nodes to detect and discard duplicate frames while identifying the transmission path in the ring topology. The tag begins with a 2-octet EtherType field set to 0x892F to denote an HSR frame, followed by three key fields: a 16-bit sequence number, a 12-bit LSDU (link service data unit) size indicating the original payload length, and a 4-bit path identifier distinguishing the frame's direction (e.g., 0 for one direction, 1 for the other).^[25] Duplicate detection relies on the combination of the 6-octet source MAC address from the Ethernet header and the 2-octet sequence number within the HSR tag. The sequence number functions as a 16-bit counter that the transmitting node (DANH or equivalent) increments by one for each successive frame sent on a given path, wrapping around after reaching 0xFFFF. Upon node reboot or power cycle, the counter resets to zero to ensure fresh identification of frames and prevent erroneous discards from prior sessions. This mechanism guarantees that each frame pair circulating the ring has a unique identifier for reliable rejection of duplicates at the receiver.^[19]^[26] The remaining 2 octets of the tag encode the LSDU size and path identifier, with the size field providing the exact length of the unmodified payload to assist in reassembly or validation, while the path identifier supports loop prevention and path-specific processing. By appending this tag, HSR frames grow by 6 octets compared to standard Ethernet, raising the maximum frame size to 1524 octets for a typical 1500-octet payload. For legacy compatibility in mixed environments, where non-HSR devices might discard oversized frames, the standard recommends configuring the network MTU to 1528 octets to accommodate the tag without fragmentation.^[36]

Compatibility with Ethernet Frames

High-availability Seamless Redundancy (HSR) integrates with standard Ethernet frames by appending a 6-byte HSR tag immediately after the source MAC address, effectively shifting the position of the subsequent fields such as the EtherType and payload while preserving the overall Ethernet frame format defined in IEEE 802.3.^[19] This modification enables duplicate detection and path identification without altering the core Ethernet header structure, ensuring that HSR-tagged frames remain compatible with Ethernet switching infrastructure that supports extended frame sizes. The HSR tag itself is identified by the EtherType value 0x892F in the current IEC 62439-3 standard, distinguishing it from regular Ethernet traffic. Additionally, HSR fully supports IEEE 802.1Q VLAN tagging, where the HSR tag follows the VLAN tag if present, allowing seamless operation in VLAN-segmented networks without requiring reconfiguration of existing tagging mechanisms.^[19] To maintain compatibility with legacy non-HSR devices, which lack the capability to process HSR tags, Redundancy Boxes (RedBoxes) are employed as intermediaries at network endpoints. These devices transparently strip the HSR tag from incoming frames before forwarding them to the legacy endpoint and insert a new HSR tag on outgoing frames from the endpoint, ensuring that the redundancy protocol operates without disrupting standard Ethernet communication at the application level.^[37] This approach allows legacy systems to participate in an HSR ring as virtual doubly attached nodes (VDANs), emulating the behavior of HSR-capable doubly attached nodes for high availability while preserving frame integrity and preventing exposure of protocol-specific elements to unsupported hardware.^[19] HSR operates directly over the standard Ethernet physical (PHY) and media access control (MAC) layers, requiring no modifications to higher-layer protocols such as IP or TCP, which remain unaware of the redundancy mechanisms.^[6] This layer-2 integration facilitates drop-in compatibility with existing Ethernet infrastructure, where HSR nodes can coexist with conventional switches and devices in mixed environments. In modern implementations compliant with IEC 62439-3, the maximum frame size is adjusted to 1528 octets to accommodate the added HSR tag alongside a full 1500-byte payload and optional VLAN tags, avoiding fragmentation and ensuring end-to-end compatibility without exceeding typical limits in industrial networks.^[36]

Timing and Synchronization

Clock Synchronization Protocols

High-availability Seamless Redundancy (HSR) networks rely on precise clock synchronization to ensure coordinated operations across nodes, with the Precision Time Protocol (PTP), defined in IEEE 1588, serving as the primary standard. As specified in IEC 62439-3 Annex C, PTP is adapted for high-availability automation networks, providing industry profiles that enable seamless time distribution over redundant Ethernet topologies like HSR. This annex outlines PTP profiles such as L3E2E (layer 3, end-to-end delay measurement) and L2P2P (layer 2, peer-to-peer delay measurement) to support deterministic synchronization in industrial environments.^[1] The protocol achieves synchronization accuracy of 1 μs, sufficient for applications like drive control and power system protection, even in a ring topology spanning up to 16 HSR nodes.^[38] This precision is maintained through PTP's master-slave hierarchy, where a grandmaster clock is elected based on factors like clock quality, priority, and stability, ensuring a single time source propagates corrections across the network. In HSR implementations, PTP messages are transported within standard Ethernet frames augmented with HSR tags, but these tags are stripped upon reception to avoid duplication and ensure single-instance processing at each node.^[19] For power utility applications, the PTP profile in IEC 62439-3 Annex C has been harmonized and adopted as IEC/IEEE 61850-9-3, which specifies sub-microsecond accuracy tailored to substation automation while integrating with redundancy protocols like HSR.^[14] This standard mandates performance requirements for grandmasters, boundary clocks, and transparent clocks to deliver reliable synchronization in fault-tolerant networks.^[38] While PTP is the normative protocol for HSR, some deployments incorporate alternatives like IRIG-B for analog time distribution or GPS for primary reference clocks, though these lack the network-integrated seamlessness of PTP and are not specified in the core HSR standard.^[38]

Synchronization Impact on HSR

Clock synchronization plays a pivotal role in High-availability Seamless Redundancy (HSR) by ensuring ordered frame delivery through precise timestamping, which is essential for event correlation in time-sensitive applications such as synchronized drives in industrial automation. In HSR networks, where frames are duplicated and transmitted along dual paths in a ring topology, synchronized clocks enable nodes to process and sequence received frames accurately, preventing misordering that could disrupt coordinated operations. This integration of time synchronization enhances the overall redundancy mechanism by providing a temporal reference that complements HSR's sequence number-based duplicate detection.^[6] Synchronized clocks significantly influence failover performance in HSR by facilitating rapid fault detection and consistent path selection across nodes. During a link or node failure, aligned timestamps allow for immediate identification of discrepancies in frame arrival times, enabling seamless switching to the alternate path without introducing timing-induced duplicates or delays. For instance, Precision Time Protocol (PTP) messages in HSR carry HSR tags and are subject to the duplicate discard logic, with only the first arriving copy processed at each node, aiding in maintaining synchronization continuity. This approach ensures that redundancy operations remain robust even under fault conditions, minimizing disruptions in high-availability environments.^[19]^[25] Integrating clock synchronization with HSR presents challenges, particularly in handling PTP delay requests over the dual redundant paths, where peer-to-peer delay measurements must be consistently calculated on each link to avoid asymmetries. In large HSR rings, synchronization accuracy can degrade due to accumulated path variations; for example, beyond 16 nodes, offsets may exceed 1 μs, impacting the precision required for applications like substation automation. To mitigate this, identical PTP configurations across ports are recommended, and redundant synchronization messages are utilized to refine clock adjustments rather than being discarded.^[39]^[40] The benefits of synchronization in HSR are pronounced in IEC 61850 environments for power system protection, where sub-microsecond accuracy enables precise event timestamping and coordinated relay operations across redundant networks. By leveraging PTP's power profile over HSR, systems achieve enhanced reliability for fault-tolerant communication, supporting zero-recovery-time redundancy while maintaining the temporal precision necessary for protective functions. This synergy ensures that HSR not only provides seamless failover but also upholds the timing integrity critical for mission-critical utilities.^[40]

Performance Characteristics

Bandwidth and Latency Metrics

In High-availability Seamless Redundancy (HSR) networks, the duplication of every Ethernet frame sent in both directions around the ring topology results in approximately doubled traffic load, reducing the effective network capacity to about 50% of the link speed for multicast and broadcast frames, as redundant copies circulate the entire ring before being discarded.^[20]^[41] For unicast traffic, the effective capacity approaches 100% at the receiver due to discard of the second arriving duplicate upon detection via the HSR tag's sequence number, though the network still experiences the full doubled load until the copy reaches the destination.^[19] This bandwidth consumption maintains nominal link speeds, such as 100 Mbit/s full-duplex, but halves the usable throughput for redundancy-enabled traffic in steady-state operation.^[6] The HSR protocol introduces a 6-octet tag inserted between the MAC header and payload of each data frame, consisting of a 16-bit EtherType (0x88FB), 4-bit path identifier, 12-bit frame size, and 16-bit sequence number for duplicate detection, adding approximately 0.4% overhead relative to typical Ethernet frame sizes (e.g., 6 bytes on a 1500-byte payload frame). Supervision frames, transmitted periodically by each node (default interval of 3 ms, configurable from 0 to 65,535 ms) to maintain ring supervision and detect topology changes, consume negligible additional bandwidth, typically less than 1% under normal conditions due to their small size and infrequency. This tag enables compatibility with standard Ethernet MTU up to 1998 bytes including overhead, without requiring frame fragmentation in most industrial applications. Latency in HSR networks benefits from cut-through forwarding, where frames are transmitted upon header receipt without full buffering, achieving end-to-end latencies of approximately 0.25 ms for 1500-byte frames in small rings (e.g., 1-2 hops at 100 Mbit/s), as measured in reference implementations.^[6] In larger configurations supporting up to 50 nodes, real-world benchmarks using RedBoxes for legacy device integration demonstrate maximum latencies of 0.5 ms for 1500-byte frames over multi-hop paths under full load, including transmit/receive processing (0.03-0.17 ms per direction) and propagation, while maintaining 100% frame delivery.^[6] These metrics ensure HSR meets real-time requirements in industrial settings, with the duplicated paths providing redundancy without introducing variable jitter beyond standard Ethernet bounds.^[6]

Fault Recovery and Efficiency

High-availability Seamless Redundancy (HSR) ensures fault recovery through a duplex transmission mechanism where each frame is simultaneously sent over two independent paths in a ring topology, allowing the receiving node to select the first arriving duplicate while discarding the second. This design results in a seamless switchover time of 0 μs, as the alternate path remains continuously active without requiring any reconfiguration or protocol reconvergence upon failure of the primary path.^[6]^[9] Fault detection in HSR relies on supervision frames transmitted at regular intervals from each node to monitor path integrity and node liveness. These frames, defined in IEC 62439-3, are sent every 3 ms by default, enabling the network to identify link or node failures in under 3 ms by detecting the absence of expected supervision messages.^[34] This rapid detection, combined with the inherent parallelism, eliminates the convergence delays typical in protocols like RSTP, which can take 50 ms or more, thereby maintaining uninterrupted operation during faults.^[42] HSR's efficiency stems from its protocol-independent operation, which avoids overhead from spanning tree algorithms or rerouting computations, ensuring deterministic performance without convergence periods. The 2021 edition of IEC 62439-3 enhances HSR efficiency in meshed configurations and with time-sensitive networking (TSN), reducing overhead in complex topologies.^[1] However, this comes with trade-offs, including approximately doubled cabling requirements due to the dual-port ring architecture and increased power consumption from simultaneous transmission on both ports. Additionally, bandwidth utilization doubles because of frame duplication, which can strain resources in high-traffic scenarios.^[2]^[20] A key limitation of HSR is its tolerance for only single-point failures, such as one link or node outage; multiple concurrent failures can disrupt the ring, necessitating hybrid configurations with Parallel Redundancy Protocol (PRP) to enhance resilience across topologies. Efficiency also diminishes in environments with asymmetric traffic patterns, where unidirectional flows overload one direction of the ring, leading to potential bottlenecks despite the redundancy.^[43] In practical deployments, such as smart grid applications, HSR has demonstrated exceptional reliability, achieving 99.999% availability in digital substation networks by providing zero packet loss during faults. For instance, Lanner Electronics' implementations as of 2025, such as the ICS-P770 appliance, support HSR/PRP in substations for real-time monitoring with instantaneous failover, aligning with IEC 61850 standards for mission-critical power systems.^[44]^[45]

Implementation Aspects

Hardware and Software Requirements

High-availability Seamless Redundancy (HSR) requires specific hardware configurations to ensure seamless redundancy without interrupting data flow. HSR nodes, known as Doubly Attached Nodes for HSR (DANH), must incorporate at least two Ethernet ports to connect to the redundant ring topology, enabling simultaneous transmission and reception of duplicated frames along both paths. These ports typically support triple-speed Ethernet operation at 10 Mbps, 100 Mbps, and 1000 Mbps to accommodate various industrial network speeds while maintaining compatibility with standard Ethernet infrastructure.^[46] To achieve the low-latency requirements of HSR, hardware implementations rely on field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs) for frame forwarding and duplicate discarding within microseconds. This hardware acceleration is essential because software-based processing cannot meet the protocol's demand for near-zero recovery time in the event of a link or node failure. Cut-through switching is employed in these devices, allowing frames to be forwarded port-to-port with minimal delay—typically in the microsecond range—by beginning transmission before the entire frame is received, thus reducing overall network latency.^[24] On the software side, HSR-capable devices need firmware that implements the duplicate discard logic as defined in IEC 62439-3 Clause 5, where each node identifies and removes redundant frames using the HSR tag's sequence number and source address to prevent loops and ensure single-frame delivery to the application layer. Linux kernel support for HSR is provided through the hsr module, which handles frame duplication, tagging, and discarding in software; recent updates in 2024 introduced offload capabilities for specific hardware like TI's ICSSG, improving stability and reducing CPU overhead for high-throughput scenarios.^[47]^[48] Processing requirements for HSR include dedicated resources for sequence number tracking, as the protocol uses a 16-bit sequence number field in the HSR tag, allowing up to 65,536 unique entries per source MAC address to detect and discard duplicates within a sliding window. This necessitates sufficient on-chip memory—often in the range of kilobytes to megabytes depending on the number of tracked sources—to store recent sequence histories without introducing delays.^[49]^[50] Reference designs facilitate HSR implementation; for instance, Texas Instruments' 2017 Ethernet reference design demonstrates HSR on Sitara processors using PRU-ICSS for low-latency redundancy in substation automation, remaining relevant for current deployments. Similarly, Microchip's 2020 application note for the KSZ9477 switch details HSR configuration on a seven-port Gigabit Ethernet device, providing guidance on tag insertion, duplicate handling, and ring integration.^[51]^[7]

Vendor Support and Deployments

Several major vendors provide hardware and software support for High-availability Seamless Redundancy (HSR), enabling its integration into industrial and utility networks. Cisco implements HSR and PRP modes in its Industrial Ethernet (IE) series switches, including the IE 4000, IE 4010, and IE 9300 models, with updated configurations documented as of September 2025 for ring topology redundancy in industrial environments.^[19]^[34] Belden offers HSR support through its Hirschmann switches and embedded modules, focusing on media redundancy for uninterrupted data communication in industrial automation.^[20] TTTech, via its Flexibilis subsidiary, provides industrial-grade HSR solutions, including the Flexibilis Ethernet Switch (FES-HSR) that has supported Gigabit-speed HSR since 2011 and undergoes regular interoperability testing.^[52]^[53] Software ecosystems further enhance HSR adoption, particularly in open-source and control systems. The Linux kernel includes an HSR driver, with developments in 2024 emphasizing improvements for embedded systems, as presented at the Linux Plumbers Conference in September 2024, to support real-time redundancy in resource-constrained devices.^[18] HSR integrates with SCADA platforms for substation automation, where protocols like PRP and HSR ensure quick failover in critical operations, as implemented in systems from vendors like COPA-DATA and Honeywell for power grid monitoring.^[54]^[55] Real-world deployments highlight HSR's role in mission-critical infrastructure. ABB's Relion 615 series protection relays incorporate HSR alongside PRP for high-availability communication in power distribution, ensuring zero frame loss in ring topologies.^[56] In smart grid networks, HSR/PRP combinations provide 24/7 uptime by adding redundancy to Ethernet infrastructures, minimizing downtimes in utility automation as deployed by TTTech solutions.^[2] Addressing deployment challenges, interoperability testing has been crucial for multi-vendor environments. Reports from Automation.com detail successful HSR/PRP tests involving Flexibilis switches and other devices, confirming compatibility in ring and parallel topologies since 2011.^[17] Cost reductions are achieved through IP cores like SOC-E's HSR-PRP Switch, a hardware-accelerated solution implementable on low-cost FPGAs for electric market devices, offering zero recovery time without specialized ASICs.^[57]^[58]

References

[1]
IEC 62439-3:2021
Dec 15, 2021 · IEC 62439-3:2021 specifies two related redundancy protocols that, in the event of failure of any network element, provide seamless switchover with zero ...
[2]
HSR/PRP for smart grid networks - TTTECH
High-availability Seamless Redundancy (HSR) is a standard (IEC 62439-3 Clause 5) that provides redundant Ethernet. It is suitable for applications that require ...
[3]
A brief introduction to High Availability Seamless Redundancy (HSR ...
High-availability Seamless Redundancy (HSR) is a standard (IEC 62439-3 Clause 5 [1]) which answers to the need for reliable Ethernet.Missing: explanation | Show results with:explanation
[4]
IEC 62439-3:2010
Feb 25, 2010 · IEC 62439-3:2009 specifies two redundancy protocols based on the duplication of the LAN, resp. duplication of the transmitted information, ...
[5]
PRP and HSR version 1 (IEC 62439-3 Ed.2), improvements and a ...
This paper presents the most remarkable improvements included, others susceptible of being included and a software prototype to be run in PCs and/or FPGAs.
[6]
[PDF] High-Availability Seamless Redundancy (HSR) Ethernet for ...
Jan 8, 2016 · HSR is a redundancy protocol for Ethernet networks, standardized as IEC 62439-3 Clause 5, and is selected as one of the redundancy protocols for ...
[7]
[PDF] AN3474 - Microchip Technology
May 20, 2020 · High-availability Seamless Redundancy is a standard defined in IEC 62439-3 Clause 5 targeting applications that require short reaction time and ...<|control11|><|separator|>
[8]
https://ieeexplore.ieee.org/document/6038906
[9]
HSR — High-availability Seamless Redundancy - iGrid T&D
HSR (High-availability Seamless Redundancy) is a redundancy protocol for Ethernet networks requiring short reaction times and high availability.Missing: explanation | Show results with:explanation
[10]
(PDF) Resilience technologies in Ethernet - Academia.edu
Unlike PRP, HASAR requires only one network 10.2. MRRT and confined to a ring topology. It is used for cost sensitive applications demanding bumpless redundancy ...
[11]
ABB-Siemens - Interoperability (PRP & HSR) Demo at Cigré 2010 ...
Several companies are developing components with the support of the KTI project of ZHAW. © Slides by HSR project team Interoperability demo at CIGRE 2010.
[12]
IEC 62439-3:2016 ED3 - Standard Norge | standard.no
Mar 31, 2016 · IEC 62439-3:2016 is applicable to high-availability automation networks based on the Ethernet technology. This part of IEC 62439 specifies two ...
[13]
IEC 62439-3:2021/COR1:2023
Apr 21, 2023 · International Standards facilitate technical innovation, efficient and sustainable energy access, smart urbanization and transportation systems, ...
[14]
IEC/IEEE 61850-9-3:2016
May 31, 2016 · IEC/IEEE 61850-9-3:2016(E) specifies a precision time protocol (PTP) profile of IEC 61588:2009 / IEEE Std 1588-2008 applicable to power utility automation.
[15]
The High-Availability Seamless redundancy protocol (HSR): Robust ...
Jul 19, 2012 · The High-Availability Seamless redundancy protocol (HSR): Robust fault-tolerant networking and loop prevention through duplicate discard.
[16]
https://www.cisco.com/c/en/us/td/docs/switches/lan/cisco_ie3x00/software/17_4/b_redundancy_17-4_iot_switch_cg/m_hsr_iosxe_iotswitch.html
[17]
HSR/PRP Redundant Ethernet Interoperability Testing conducted at ...
Flexibilis Ethernet Switch with High-availability Seamless Redundancy and Parallel Redundancy Protocol support (FES-HSR) is a triple-speed (10Mbps/100Mbps/1Gbps) ...
[18]
High-availability Seamless Redundancy (HSR) - introduction ...
High-availability Seamless Redundancy (HSR) - introduction, current Linux driver status and further development. 18 Sept 2024, 11:00.
[19]
Chapter: High-Availability Seamless Redundancy (HSR) - Cisco
Sep 23, 2025 · HSR is similar to Parallel Redundancy Protocol (PRP) but is designed to work in a ring topology. Instead of two parallel independent networks of ...Missing: explanation | Show results with:explanation
[20]
High Availability Seamless Redundancy - Belden
HSR functions primarily as a protocol for creating media redundancy and can be built with genuinely uninterrupted data communication.Missing: explanation | Show results with:explanation
[21]
High-Availability Seamless Redundancy | FortiSwitch 7.6.4
HSR provides seamless communication with fault tolerance by duplicating every unicast frame sent in HSR networks.Configuring the HSR settings · Configuring the HSR ring · Troubleshooting HSRMissing: explanation | Show results with:explanation
[22]
Cost-Effective Topology Design for HSR Resilient Mesh Networks
The IEC 62439-3 recommendation [8] introduces two Ethernet-based redundancy methods sharing the capability of zero failover time, namely, the parallel ...
[23]
[PDF] Redundancy Protocol Configuration Guide, Cisco Catalyst IE3x00 ...
Aug 10, 2020 · The switch supports two HSR modes of operation: HSR-SAN and HSR-PRP. HSR-PRP mode is added in the Cisco IOS XE 17.14.1 release. The switch ...
[24]
[PDF] Industrial Utilities Resiliency - PRP and HSR Best ... - Cisco Live
IEC 61850 standards define a model for critical traffic flows. • GOOSE and SV traffic flows require the highest priority with the lowest latency, among others.
[25]
HSR/PRP - WeOS User Guides
High-Availability Seamless Redundancy (HSR) and Parallel Redundancy Protocol (PRP), defined in IEC 62439-3, provides seamless failover on link failure.Missing: structure DAN
[26]
[PDF] IEC 62439-3 HSR/PRP Implementation on Sitara Processors using ...
Each node on the network has two network ports and is directly inserted in a network with ring topology. • Supports 10/100 mbps full-duplex Ethernet interface.
[27]
[PDF] SIPROTEC with HSR and PRP - Profile - Siemens
▫ Cut through switching in HSR to ensure low Latency. ▫ Use Redundant Substation Controller. ▫ Symmetrical and fully separated PRP-A and PRP-B networks.<|control11|><|separator|>
[28]
[PDF] IEC 62439-3 - iTeh Standards
It specifies the PRP and HSR seamless redundancy protocols. It was adopted by IEC TC57 WG10 as the redundancy method for demanding substation automation ...
[29]
Understanding PRP and HSR in IEC 61850 Substation Networks
High-availability Seamless Redundancy (HSR) is a protocol also defined in IEC 62439-3. Like PRP, it provides zero-time recovery in case of a network failure—but ...
[30]
https://ieeexplore.ieee.org/document/5347037
[31]
(PDF) Seamless and low-cost redundancy for substation automation ...
HSR is specifically designed for applications requiring high availability and zero downtime, providing seamless redundancy by creating duplicate paths for data ...Missing: development Alstom RuggedCom
[32]
HSR - CloudEngine S3700, S5700, and S6700 V600R025C00 Web ...
High-availability Seamless Redundancy (HSR) is an ... Mode U: This mode is similar to mode H, but ... Mode X: This mode is similar to mode H, but ...
[33]
High-availability Seamless Redundancy - Wikipedia
High-availability Seamless Redundancy (HSR) is a network protocol for Ethernet that provides seamless failover against failure of any single network component.
[34]
IEC 62439-3.5 HSR Kirrmann | PDF | Computers - Scribd
HSR - High-availability HSR principle (mode U, unicast frames removal) ... HSR - High-availability HSR principle (mode X, multicast frame removal) Seamless ...
[35]
Introduction - iS5 Communications
HSR Frame. Figure 2. PRP Frame. Periodically, supervision frames allow supervision of the status of the redundant network, e.g. broken links can be detected.
[36]
Redundancy Protocol Configuration Guide, Cisco Catalyst IE9300 ...
Aug 8, 2025 · When CDP or LLDP is enabled, you can use the CDP or LLDP information to find the adjacent nodes on an HSR ring and their status. You can then ...
[37]
PROCESSOR-SDK-AM64X: The maximum MTU of HSR interface ...
The kernel HSR subsystem limits the maximum MTU of the HSR ... 1528 bytes(14-byte Ethernet header + 4-byte VLAN ... HSR tag insertion when HSR is offloaded).
[38]
Seamless Redundancy with PRP and HSR - ZHAW
The Parallel Redundancy Protocol (PRP) and the High-Availability Seamless Redundancy Protocol (HSR) enable completely uninterrupted handling of single network ...Missing: inverters | Show results with:inverters
[39]
[PDF] Conformance Class IEC/IEEE 60802 - IEEE 802
IEC 62439 ... Frame size. Quantity. 1532. 1532. 1532. 1532. 2000. 1532 ... HSR TAG supported. Feature. Not used. Supported but.
[40]
[PDF] IEC / IEEE - National Institute of Standards and Technology
Nov 18, 2020 · The resulting standard IEC/IEEE 61850-9-3 provides sub-microsecond time accuracy and supports seamless redundancy. Precision Time. Protocol ...
[41]
High-Availability Seamless Redundancy (HSR) for IE 4000, IE 4010 ...
High-availability Seamless Redundancy (HSR) is defined in International Standard IEC 62439-3-2016 clause 5. HSR is similar to Parallel Redundancy Protocol ...Missing: objectives | Show results with:objectives
[42]
[PDF] Best Practices for Implementing PTP in the Power Industry
This paper explores the implementation of PTP across different Ethernet network architectures and looks at methods to provide PTP redundancy. In addition, the ...
[43]
https://www.westermo.com/us/industries/energy/substation-automation/hsr-prp-redundancy
[44]
High-availability Seamless Redundancy (HSR) Traffic Reduction ...
May 8, 2015 · This paper investigates several existing techniques for reducing high-availability seamless redundancy ... Hubert Kirrmann. Hubert Kirrmann. HSR - ...
[45]
https://www.researchgate.net/publication/280114681_Availability_Analysis_and_Optimization_of_HSR-based_Ethernet_Mesh_Networks
[46]
HSR and PRP in substation networks - Westermo
Presented on this page are application examples that use either PRP, HSR or a combination of both to create redundant and effective data networks.
[47]
Building Resilient Digital Substations with High-Availability HSR ...
Nov 22, 2024 · High-availability Seamless Redundancy (HSR) and Parallel Redundancy Protocol (PRP) are vital technologies that guarantee zero packet loss ...
[48]
Availability Analysis and Optimization of HSR-based Ethernet Mesh ...
Jul 19, 2015 · This article addresses a network design problem for smart grid communications. We propose a simple network design method for IEC 62439 HSR ring ...
[49]
HSR/PRP solutions for FPGAs - TTTECH
It provides the HSR/PRP supervision feature described in IEC62439-3 standard (HSR/PRP), which monitors HSR/PRP redundancy operation. If there is a single ...
[50]
4.2. HSR_PRP — Processor SDK Linux Documentation
HSR stands for High Availability Seamless Redundancy. This is a protocol used to support redundant networks needed for industrial applications.Missing: benchmark | Show results with:benchmark
[51]
Introduce HSR offload support for ICSSG - LWN.net
Sep 4, 2024 · Introduce HSR offload support for ICSSG ; Subject: [PATCH net-next v4 0/5] Introduce HSR offload support for ICSSG ; Date: Wed, 04 Sep 2024 15:35: ...
[52]
4.2. HSR_PRP — Processor SDK Linux Documentation - http
HSR and PRP are protocols for redundant network connections, using two ethernet ports. HSR is a ring, while PRP can be any configuration.
[53]
[PDF] BCM56070 Switch Programming Guide
Aug 12, 2020 · This is a programming guide for the BCM56070 switch, including an introduction, architectural overview, and device interfaces.
[54]
[PDF] High-Availability Seamless Redundancy Ethernet Reference Design ...
Oct 1, 2017 · This Ethernet reference design implements high- reliability, low-latency network communication for substation automation equipment in Smart Grid.Missing: inverters | Show results with:inverters<|separator|>
[55]
TTTECH Flexibilis
As early as 2011, Flexibilis released a version of FES supporting the HSR (High-availability Seamless Redundancy) protocol – the first one with Gigabit speed.
[56]
Flexibilis Ethernet Switch with HSR - Lattice Semiconductor
Flexibilis Ethernet Switch with High-availability Seamless Redundancy support (FES-HSR) is a triple speed (10Mbps/100Mbps/1Gbps) Ethernet Layer 2 switch.
[57]
How to achieve robust and high availability communication for ...
However, when running critical functions on a SCADA system, it's recommended to use PRP or HSR to achieve the quick network failover times needed. PRP also ...
[58]
SCADA - Honeywell Process Solutions
Honeywell provides a highly scalable, integrated, multiserver SCADA solution that incorporates innovative applications for human machine interface (HMI) to ...
[59]
[PDF] — Relion® 615 series Protection and control relays - ABB
The relays further support both the parallel redundancy protocol (PRP) and the high- availability seamless redundancy (HSR) protocol, together with the DNP3 ...
[60]
[PDF] 10M/100M/1G HSR-PRP Switch IP Core - SOC-E
It is a full hardware solution that can be implemented in different FPGA families, including low-cost versions. It is a flexible solution for the Electric ...Missing: reduction | Show results with:reduction
[61]
FPGA Ethernet switch IP cores with HSR (IEC 62439-3) support from ...
Jun 13, 2011 · The triple speed (10Mbps/100Mbps/1Gbps) FES IP cores operate on Ethernet Layer 2 and can switch with Gigabit forwarding capacity per port.Missing: dual- | Show results with:dual-