Shipping Media Over QUIC at Meta Scale for Instagram Live Ingest

This post details the deployment of Media over QUIC (MoQ) for Instagram Live ingest on iOS and Android, marking a significant application of the emerging IETF standard at Meta's scale. It describes MoQ's architecture, built on QUIC, featuring Tracks, relays for caching and fan-out, and tunable latency. The deployment strategy involved client integration using the moxygen library, an edge translation layer to interface with existing RUSH-based backends, and server-side routing for phased rollouts. Key operational hurdles addressed include evolving specifications, handling network firewalls by selecting TCP or QUIC at connection start, and developing MoQ-aware tracing for debugging encrypted QUIC traffic. The outcome achieved parity with the previous RUSH protocol, validating MoQ as a production-ready transport for large-scale live video ingest and setting the stage for its extension to playback and other real-time AI applications.
More from the Internet Latency and Network Performance Analysis trail
- OCP Summit 2026: Powering AI with Open Innovations in Scale-Up, Scale-Out, and Scale-Across Networks
- NTS: Authenticated Time at Meta
- Inside Petal: Building the World’s First Petabit-Class Transoceanic Subsea Cable
- ZGateway: Learnings from Putting a Proxy in Front of ZippyDB
- MetaRoCE: A New RDMA Transport Built for AI-Scale Ethernet