Quicly achieves 683ms latency using Media over QUIC and Cloudflare

This technical milestone demonstrates that Media over QUIC can deliver sub-second latency for complex browser-based monitoring without requiring proprietary contribution hardware. By decoupling the player session from the visual layout, Quicly addresses a common stability issue in web-based multiviews where UI changes often trigger stream restarts. This approach signals a shift toward more flexible, low-latency production tools that leverage standard CDN infrastructure like Cloudflare rather than specialized hardware relays. As the industry moves away from server-rendered composites, watch for how Bitmovin and other infrastructure providers integrate these MoQ readiness checks to ensure cross-browser playback stability.

Media over QUIC (MoQ) is gaining traction as a transport layer for low-latency video delivery, with multiple infrastructure providers building on the IETF draft specification. Cloudflare has been a key contributor to the protocol's development, publishing its MoQ relay implementation as part of its broader streaming infrastructure stack. The company's relay network allows publishers to ingest streams at the edge and distribute them to viewers with minimal hop count, which is the architecture Quicly uses for its multiview system. Cloudflare announced in early 2025 that its MoQ relay service was available for developers building real-time applications on top of its global network, positioning the protocol as a complement to its existing WebRTC and HLS delivery options.

The competitive landscape for sub-second browser delivery includes WebRTC, CMAF-LL, and SRT-based pipelines, each with different trade-offs around scalability, firewall traversal, and codec support. Bitmovin, which provides player infrastructure used in production monitoring workflows, has been tracking MoQ readiness across browsers and has published guidance on fallback strategies for environments where QUIC transport is not yet natively supported. Bitmovin's engineering team outlined in 2025 how MoQ could reduce the complexity of multi-angle live streams by eliminating the need for separate WebRTC peer connections per viewer, noting that relay-based architectures scale more predictably than peer-to-peer approaches for production monitoring use cases.

On the technical side, the 683ms median glass-to-glass latency reported by Quicly compares favorably against typical WebRTC implementations, which often range from 500ms to 1.5 seconds depending on network conditions and encoding complexity. The IETF MoQ working group has continued to refine the draft specification, with recent revisions focusing on track-level prioritization and subscription management that enable the kind of independent angle control Quicly demonstrates. The IETF MoQ working group published updated draft specifications in 2025 that introduced group-level forwarding hints and improved congestion response mechanisms, both of which are relevant to multiview scenarios where multiple simultaneous tracks must be managed within a single browser session.