ST 2110: Separate Media, Precise Timing and Redundant Networks

Why an AV engineer should care

ST 2110 is the heavyweight end of professional media networking. You will encounter it in broadcast facilities, arenas, production studios, large command centers and some high-end corporate or entertainment projects.

It uses familiar ingredients—Ethernet, IP, UDP, RTP, multicast and PTP—but expects much tighter engineering than a typical conference-room AVoIP VLAN.

The central idea is simple:

Instead of carrying one bundled “video signal,” ST 2110 carries video, audio and ancillary data as separate, precisely timed network streams.

That flexibility is powerful. It also means the network becomes part of the real-time media system rather than merely the plumbing underneath it.

Learning objectives

By the end of this lesson, you should be able to:

• Explain the separate-essence model.
• Identify the major ST 2110 documents.
• Describe how PTP and RTP timestamps keep streams aligned.
• Explain ST 2022-7 seamless redundancy.
• Recognize why ST 2110 requires serious bandwidth, timing and monitoring design.

---

1. ST 2110 is a suite

SMPTE ST 2110 defines the transport, timing and description of separate professional-media streams over managed IP networks. SMPTE ST 2110 overview. It does not define an entire finished system, controller or switch configuration.

The major pieces are:

The suite has additional parts, but these explain most systems an AV engineer will initially encounter. SMPTE continues to revise and expand the suite; for example, ST 2110-30 was revised in 2025.

What “active video” means

An SDI signal includes active image pixels plus blanking intervals, embedded audio and ancillary data.

ST 2110-20 transports the active video essence. Audio and ancillary information travel separately. It does not simply stuff the entire SDI signal into Ethernet packets.

That creates useful possibilities:

• A production switcher can receive video from a camera.
• An audio console can independently receive the camera’s audio.
• A captioning system can process its ancillary data.
• Each stream can route to different destinations.
• Audio can change without interrupting video.
• A device need only receive the essences it actually uses.

SMPTE describes these individual streams as independently timed and able to take different paths through the network.

The tradeoff is that something must keep all those pieces aligned.

---

2. PTP replaces the facility timing plant

Traditional broadcast systems distribute black burst, tri-level sync, word clock and timecode through dedicated connections.

ST 2110 uses PTP, following the SMPTE ST 2059 profile, to give network devices a highly accurate shared understanding of time. ST 2059 builds on IEEE 1588 PTP specifically for professional-media applications.

A PTP grandmaster effectively tells the facility:

This is the current time, and this is the exact rate at which time is advancing.

Media packets also carry RTP timestamps. Those timestamps connect each video frame or audio sample to the common timeline.

That allows a receiver to determine:

• When a video frame was captured
• Which audio samples belong with it
• When those essences should be played
• Whether a packet is early, late or missing

PTP therefore provides both frequency and phase alignment. NTP may put two computers within milliseconds of the same wall-clock time. ST 2110 commonly needs media devices aligned far more closely. Keep media-clock alignment separate from network packet arrival jitter; receiver buffering still matters.

PTP-aware switching

Large systems often use:

• Boundary clocks: A switch synchronizes to upstream PTP and acts as a timing source to downstream devices.
• Transparent clocks: A switch measures and reports the delay packets experienced while passing through it.
• Redundant grandmasters
• Dedicated PTP monitoring
• Carefully controlled priorities and domains

A timing fault can affect hundreds of streams simultaneously. PTP health must be monitored like a core system service, not treated as background multicast that probably sorts itself out.

---

3. Packet pacing matters

A sender could theoretically produce the correct average video bitrate while releasing packets in giant bursts.

The average might be 2.5 Gb/s, but a short burst could overflow a switch queue or receiver buffer. Packets would be dropped even though link-utilization graphs appeared reasonable.

ST 2110-21 defines video traffic-shaping and delivery-timing behavior. It controls how senders distribute packets over time and tells receivers what burst behavior to expect.

This is one reason ordinary bandwidth calculations are necessary but insufficient.

For uncompressed 10-bit 4:2:2 active video:

1920 × 1080 × 59.94 × 20
≈ 2.49 Gb/s

For 3840 × 2160 at 59.94:

3840 × 2160 × 59.94 × 20
≈ 9.94 Gb/s

Those are active-pixel calculations before Ethernet, IP, UDP and RTP overhead. A 4K60 uncompressed flow therefore does not fit through a 10 Gb/s port.

The network must handle:

• Sustained bitrate
• Packet overhead
• Packet bursts within the permitted profile
• Multiple simultaneous multicast groups
• Audio and ancillary flows
• Redundant copies
• Failure-path traffic

---

4. SDP describes the stream; NMOS operates the system

An ST 2110 receiver needs details such as:

• Destination multicast address
• UDP port
• Payload type
• Video resolution and frame rate
• Chroma sampling and bit depth
• Audio sample rate and channel arrangement
• Clock reference
• RTP timing information

These details are represented in an SDP, or Session Description Protocol, document.

You can think of SDP as the stream’s technical recipe. If the sender is producing 1080p59.94 4:2:2 10-bit video, the receiver needs the right recipe to interpret the incoming packets.

SDP by itself does not create a friendly routing system. That is where NMOS commonly enters:

• IS-04 discovers and registers resources.
• IS-05 connects senders and receivers.
• The controller presents logical sources and destinations to the operator.

ST 2110 defines media transport. NMOS helps turn hundreds of streams into an operable facility. AMWA describes the same division: ST 2110 handles streaming, while NMOS addresses discovery, connection and monitoring.

---

5. ST 2022-7: two paths, one uninterrupted stream

ST 2022-7 provides seamless protection by sending two matching RTP packet streams through different network paths.

These are commonly called:

• Red network
• Blue network

The receiver combines the duplicate packet streams, accepting a valid copy within its supported differential-delay window and discarding duplicates.

If Red loses packet 18,432 but Blue delivers it, the receiver still reconstructs a complete stream. There is no visible switch between networks because protection happens packet by packet. SMPTE describes this as hitless protection over diverse physical paths.

For this to provide real protection, the paths need actual independence:

• Separate switches
• Separate uplinks
• Separate fiber routes where practical
• Separate power sources
• Sufficient capacity on each network
• Correctly engineered delay differences
• Monitoring of both paths

One stream looking healthy can hide a failed redundant path. Receivers should report packet loss or late packets independently for Red and Blue so the degraded protection is discovered before the second failure.

ST 2022-7 protects the RTP media packets. PTP redundancy requires its own careful design; you cannot assume that duplicating media automatically makes timing resilient.

---

6. Commercial-AV examples

Example 1: Arena production system

An arena has eight cameras, replay servers, a production switcher, an audio console, recorders and several multiviewers.

Each camera gateway produces:

• One 2110-20 video essence
• Several 2110-30 audio essences
• A 2110-40 ancillary-data essence
• Matching Red and Blue copies

The audio console subscribes only to audio. The production switcher takes video and audio. Recorders capture all required essences. Multiviewers receive selected video streams at once.

The designer must calculate traffic separately for each fabric and each failure condition. Red and Blue each need enough capacity to carry the complete required system load.

Example 2: IPMX meeting ST 2110

A large corporate production facility uses ST 2110 internally, but conference rooms use IPMX endpoints for HDMI laptops, displays, HDCP and USB.

A gateway provides the boundary:

• The room side handles commercial-AV HDMI behavior.
• The production side exposes appropriate ST 2110 video and audio essences.
• NMOS advertises the gateway’s senders and receivers.
• The gateway aligns asynchronous HDMI sources to the facility’s PTP timing.

The designer should document the gateway’s format conversion, latency, audio mapping, HDCP behavior and failure response. Calling both sides “AV-over-IP” does not make the boundary disappear.

---

Design documentation takeaway

For an ST 2110 design, document:

• Every video, audio and ancillary essence
• Format, resolution, frame rate and bitrate
• Sender and receiver relationships
• Multicast address and UDP port
• SDP source and management method
• PTP profile, domain and grandmaster hierarchy
• Boundary- or transparent-clock behavior
• ST 2110-21 sender capability
• Red and Blue interfaces and physical paths
• Normal and failure-path bandwidth
• NMOS registry and controller dependencies
• Monitoring for loss, late packets and PTP offset
• Tested firmware and interoperability results

JT-NM test results should be treated as a capability snapshot for the tested firmware—not a permanent universal certification. JT-NM Tested.

Practical troubleshooting exercise

A dual-network receiver shows clean video, but monitoring reports:

• Red: no packet loss
• Blue: increasing packet loss
• Output: no visible errors
• PTP: locked

Is the system healthy?

The output is currently protected because Red supplies every packet. The system is nevertheless in a degraded redundancy state. If Red fails, Blue may not sustain the stream.

Investigate Blue’s interface counters, optics, fiber path, switch queues, multicast forwarding and bandwidth. Do not wait for the picture to fail before acknowledging the problem.

Compact glossary

• Essence: Individual video, audio or data component.
• SDP: Machine-readable description of a media stream.
• ANC: Ancillary data such as captions or timecode.
• ST 2059: SMPTE timing framework and PTP profile.
• ST 2022-7: Seamless redundant RTP transport.
• Boundary clock: PTP-aware device that follows upstream timing and serves downstream devices.
• Packet pacing: Distribution of packets over time to control bursts.
• Hitless protection: Recovery from a path failure without interrupting the reconstructed media.

Review questions

1. Why can ST 2110 route a camera’s video and audio independently?
2. What problem does ST 2110-21 address?
3. Why should Blue-network errors be investigated while Red still produces clean video?

---

Answers

1. Video, audio and ancillary data travel as separate network essences.

2. It constrains video packet timing and burst behavior so senders do not overwhelm switches or receiver buffers.

3. The system has lost reliable redundancy. A later Red-network failure could interrupt the media.

Where this sits in the course

We now have the major AVoIP families in view: Dante/AES67, NDI, IPMX and ST 2110.

Lesson 24 will turn these standards into a practical network-design workflow: requirements gathering, traffic matrices, VLANs, multicast, uplink sizing, redundancy, switch selection, drawings, commissioning records and the questions the project team should settle before anybody orders equipment.