Ensuring the reliability of critical AES67 audio streams is a top priority in professional audio environments such as live broadcast, recording studios, and corporate AV systems. When network failures or disruptions occur, the loss of audio data can lead to missed deadlines, degraded listener experience, or even safety incidents. Creating redundant paths for these streams provides a fail-safe mechanism that maintains continuous audio delivery. This expanded guide dives deeper into the technical aspects of AES67 redundancy, offering practical steps, configuration details, and best practices to prevent data loss and ensure uninterrupted operation.

Understanding AES67 Redundancy

AES67 is an open standard developed by the Audio Engineering Society for high-performance audio-over-IP (AoIP) interoperability. It ensures that devices from different manufacturers can transmit and receive uncompressed, low-latency audio over standard IP networks. Redundancy in AES67 involves creating multiple network paths so that if one path fails—due to a faulty cable, switch failure, or congestion—the audio stream can seamlessly switch to an alternate route with minimal or no glitch. This is achieved through two primary methods: stream redundancy (duplicate audio packets sent over separate paths) and path redundancy (redundant network infrastructure and routing). Both approaches work together to protect against single points of failure.

To fully understand how redundancy functions, it’s helpful to know the underlying transport protocols. AES67 uses RTP (Real-time Transport Protocol) over UDP/IP for audio data, with RTCP (RTP Control Protocol) for stream synchronisation and monitoring. For redundancy, the ST 2022-7 standard (Seamless Protection Switching) is commonly employed, allowing two identical streams to be sent on different paths—the receiver selects the best available data. This principle is central to reliable AES67 deployments. For more on the AES67 standard itself, refer to the official AES67 document.

Network Infrastructure Planning for Redundancy

Successful AES67 redundancy begins with a robust network design. A few key infrastructure choices dramatically affect reliability:

  • Redundant switches and links: Use at least two core switches connected by redundant uplinks. Each AES67 device should have dual Ethernet ports connecting to separate switches to avoid a single switch failure taking down all streams.
  • Spanning Tree Protocol (STP) and its variants: STP prevents loops in redundant networks by blocking redundant paths until a primary path fails. For low-latency audio, use Rapid Spanning Tree Protocol (RSTP - IEEE 802.1w) or Multiple Spanning Tree Protocol (MSTP) to reduce failover times to a few seconds. In some cases, Media Redundancy Protocol (MRP - IEC 62439-2) is preferred for sub-10ms failover in deterministic networks.
  • Quality of Service (QoS): AES67 streams require prioritisation to avoid jitter and packet loss. Configure DSCP (Differentiated Services Code Point) markings to ensure audio packets are queued ahead of bulk data.
  • Network segmentation: Isolate audio traffic on a dedicated VLAN with sufficient bandwidth (typically 100 Mbps or 1 Gbps per stream) to minimise interference from other data.

For more detailed network best practices, see the Dante and AES67 integration guide from Audinate, which covers VLAN design and QoS settings.

Configuring Dual Network Interfaces on AES67 Devices

Most professional AES67 devices (e.g., amplifiers, microphones, mixing consoles) come with dual Ethernet ports. Configuring them for redundancy typically involves setting both interfaces with separate IP addresses on different subnets or VLANs, connected to physically separate network paths.

  • Interface bonding vs. active/standby: Some devices support link aggregation (LACP) for bandwidth but not redundancy. For AoIP, an active/standby configuration is better: one interface is primary, the other is backup. The device monitors the primary link; if it fails, it switches to the backup.
  • Stream mapping: In the device’s web interface or management software, assign each audio stream to a specific interface. For ST 2022-7 redundancy, configure two independent stream instances (each with unique SSRC and multicast addresses) on separate interfaces.
  • Duplicate stream transmission: When using RTP redundancy, the encoder sends two identical copies of each audio packet but with different RTP sequence numbers or timestamps so the receiver can merge or switch seamlessly.

For a concrete example, refer to the Merlijn van Veen study on AES67 seamless redundancy which provides configuration examples for common hardware.

Implementing Redundant Routing Protocols

Beyond device interfaces, layer 3 routing protocols ensure that if a router or network path fails, traffic is rerouted without manual intervention. Key protocols for AES67 redundancy include:

  • Virtual Router Redundancy Protocol (VRRP): Creates a virtual IP address shared between two routers. One router is master, the other is backup. If the master fails, the backup takes over transparently. VRRP failover typically occurs within a few seconds.
  • Hot Standby Router Protocol (HSRP): Cisco’s proprietary alternative to VRRP with similar operation.
  • Equal-Cost Multi-Path (ECMP) routing: Distributes traffic across multiple active paths. For redundancy, ensure that if one path fails, traffic shifts to another via dynamic routing protocols like OSPF or BGP.
  • PIM (Protocol Independent Multicast) redundancy: AES67 streams often use multicast. Configure PIM-SM with multiple rendezvous points or use PIM-BIDIR for fault-tolerant multicast forwarding.

Documentation for VRRP and HSRP is widely available; consult the Cisco HSRP configuration guide for detailed setup steps.

Setting Up Stream Redundancy with RTP/RTCP and ST 2022-7

The most widely adopted method for AES67 stream redundancy is the ST 2022-7 standard (Seamless Protection Switching for IP Networks). It enables a receiver to accept two identical RTP streams from different network paths and combine them into a single output without gaps or duplicates.

  • Configuration on the transmitter: Configure the device to send two duplicate streams with the same payload type, clock rate, and audio format. Assign different source IP addresses and port numbers for each stream. The RTP sequence numbers and timestamps should be identical for both streams (or follow a predictable difference).
  • Configuration on the receiver: Enable ST 2022-7 mode. The receiver listens on two ports simultaneously. It uses the stream with the lowest packet loss rate or highest sequence number integrity. In the event of a complete path failure, the switchover is sample-accurate and imperceptible.
  • Clock synchronisation: Both streams must be synchronised to the same PTP (Precision Time Protocol) grandmaster clock to avoid time offsets. AES67 relies on IEEE 1588-2008 (PTPv2) for precise timing. Ensure all switches are PTP-aware and operate as transparent clocks or boundary clocks.

For a deeper dive into ST 2022-7, the SMPTE ST 2022-7 official page provides the full specification.

Leveraging Redundancy Management Software

Manual configuration is error-prone; redundancy management software automates monitoring and failover decisions. Many AES67 systems integrate with tools like Dante Controller (for Dante/AES67 hybrid networks), WheatNet-IP, Q-SYS, or standalone solutions like PRISM Sound’s Redundancy Manager. These tools offer:

  • Real-time health monitoring: Display packet loss, jitter, and latency for each stream. Alerts when thresholds are exceeded.
  • Automatic switchover: If a stream degrades beyond a set threshold, the software can seamlessly shift to the redundant stream or reroute the primary stream to a backup path.
  • Logging and analytics: Record failures for post-event analysis, helping identify recurring issues in network infrastructure.
  • Centralised configuration: Manage redundancy for hundreds of devices from a single interface, ensuring consistent settings across the system.

For those using Dante-based AES67, Audinate’s Dante Controller provides a free, powerful tool for redundancy management and stream monitoring.

Best Practices for Reliable Redundancy

Implementing redundancy is not enough; long-term reliability requires ongoing practices:

  • Network segmentation and VLAN isolation: Keep AES67 traffic on its own VLAN to prevent broadcast storms from other services. Use IGMP snooping and querier to manage multicast groups efficiently.
  • Regular testing of failover mechanisms: Schedule quarterly tests where you physically disconnect a cable or power down a switch to verify automatic switchover. Document fallover times.
  • Monitoring and alerts: Use SNMP-based network management systems (e.g., PRTG, Zabbix) to track switch CPU utilisation, port errors, and PTP clock health. Configure email/SMS alerts for critical failures.
  • Documentation: Maintain up-to-date network diagrams showing cable paths, switch configurations, IP addressing, and VLAN assignments. Include screenshots of stream redundancy settings. This is invaluable for troubleshooting.
  • Firmware updates: Keep all network switches and AES67 devices on the latest manufacturer firmware to patch vulnerabilities and improve STP convergence times.
  • Capacity planning: Ensure network bandwidth is sufficient for duplicate streams. Each stereo AES67 stream at 48 kHz/24-bit uses about 3 Mbps. Two streams = 6 Mbps per channel. Multiply by number of channels and plan headroom.

Testing and Validation

Before deploying redundancy into a production environment, rigorous testing is essential:

  • Lab simulation: Create a test network with the same switch models and cable lengths as the production setup. Introduce faults by disconnecting links or generating traffic floods with iperf.
  • Packet capture analysis: Use Wireshark on the receiver side to verify that duplicate RTP packets arrive with expected sequence numbers and timestamps. Check for gaps or out-of-order packets during failover.
  • Audio quality assessment: Record the output during failover events and listen for clicks, pops, or silence. Use tools like GL Tools or EA Acoustics to measure loss-of-audio duration.
  • Stress testing with high load: Simulate a conference with dozens of active streams to verify that redundancy remains stable under heavy network utilisation.
  • PTP performance: Monitor PTP offset and delay during failover to ensure clock synchronisation remains within AES67 tolerance (±1 μs relative to the grandmaster).

For a comprehensive testing methodology, the Audinate Dante Redundancy Testing Guide provides practical steps applicable to AES67 networks.

Conclusion

Creating redundant paths for AES67 audio streams is a critical investment for any professional audio environment where data loss is unacceptable. By designing a resilient network infrastructure—with dual switches, RSTP or MRP, VRRP routing, and ST 2022-7 stream redundancy—you can achieve sub-millisecond failover times and virtually eliminate audio dropouts. Proper configuration of dual interfaces, regular testing, and robust monitoring complete the solution. While the initial setup requires careful planning and investment, the payoff is uninterrupted, high-quality audio delivery that meets the demands of live broadcast, event production, and mission-critical communications. Adopting these practices will help you build an AES67 system that is as reliable as it is powerful.