Media over QUIC (MoQ) in Action: Multi-Track, Low-Latency Streaming

Natalia Gonzalez
11 hours ago
3 min read

Modern robotics, edge AI, and embedded vision systems are pushing traditional streaming technologies to their limits. When a robot carries multiple cameras and sensors, every millisecond counts—and every data stream matters.

That’s where Media over QUIC (MoQ) comes in.

At RidgeRun, we are actively advancing MoQ-based streaming for embedded systems, and our latest demo showcases a real-world scenario with multi-track, low-latency video streaming plus synchronized metadata over QUIC.

Let’s break down why this matters—and how we’re leading the way.

What is Media over QUIC (MoQ) and Why Does It Matter?

Media over QUIC (MoQ) is an emerging media delivery architecture built on top of the QUIC transport protocol. It is designed specifically for real-time, low-latency, and multi-source media distribution.

Traditional streaming protocols were primarily designed for single video sessions delivered from a server to many viewers. Modern systems, however — especially in robotics, edge AI, and embedded vision — generate multiple concurrent media sources, often combined with telemetry and metadata. MoQ addresses this evolution directly.

Built on QUIC

QUIC is a secure, UDP-based transport protocol that powers HTTP/3. It provides:

Encrypted connections by default (TLS 1.3)
Stream multiplexing within a single connection
No head-of-line blocking between streams
Faster connection setup
Improved congestion control

These transport-level advantages make QUIC an ideal foundation for next-generation media delivery.

A Track-Based Media Model

What makes MoQ fundamentally different is its media-aware structure.

Instead of treating media as one continuous stream, MoQ organizes content into:

Channels – logical groupings of related media
Tracks – independent media streams within a channel

Each track can represent:

A camera feed
An audio stream
A telemetry stream
Metadata associated with video
Any other time-based media source

Clients can dynamically subscribe to individual tracks within a channel, enabling selective reception of only the media they need.

This model is particularly powerful in systems that produce multiple simultaneous data streams.

MoQ Architecture Overview

Media over QUIC uses a relay-based architecture to balance scalability and latency. Instead of forcing direct peer-to-peer connections, MoQ introduces relays that form a distributed media delivery network.

The main roles in a MoQ system are:

Publisher: Produces and publishes media content.
Subscriber: Consumes media content.
Relay: Routes and forwards media between publishers and subscribers.

Publishers and subscribers do not connect directly to each other. Instead, each connects to a local relay, which participates in a larger relay network.

RidgeRun's Gst-MoQ Library

To demonstrate MoQ’s capabilities in a realistic environment, we built a MoQ-based streaming system consisting of:

A MoQ Publisher Application
A MoQ Player Application
Components of the RidgeRun MoQ plugin

MoQ in Action:

https://www.youtube.com/watch?v=BpUWY7ngBEM

MoQ Publisher Application

The publisher simulates a robot equipped with multiple cameras and telemetry systems.

Key features:

Configurable track count: Publish multiple video tracks within a single MoQ channel.
At least 4 concurrent video tracks: Each representing a different camera on the robot.
SEI metadata injection on one track: Video stream embeds valid SEI metadata.
Synthetic telemetry module: Simulates robot data such as:
- Cam Resolution
- CPU Temperature
- CPU Usage
Stable long-duration operation: Designed to run continuously for ≥1 hour without interruption.
Consistent metadata updates: SEI metadata is updated at a configurable rate, without stream disruption.