How multi-CDN works
Decisions are driven by measurements: synthetic tests, real-user data from players, or both.
- DNS-based switching: the stream's hostname resolves to different CDNs according to region or measured performance
- Manifest-based switching: the manifest lists more than one CDN, and the player chooses or is told which to use
- Mid-stream switching: the player moves to another CDN during playback if segments slow down or fail
Why services use it
- Resilience: an outage or congestion on one network does not take the service down
- Regional performance: no CDN is the fastest everywhere
- Capacity: very large live events can exceed what one network will commit to
- Commercial leverage: traffic can be shifted toward the better price
What it costs
Each CDN must be configured identically for caching, tokens and geo rules, and kept that way. Traffic split between networks earns smaller volume discounts on each, and each must fill its own cache from the origin. The switching layer itself is another system to buy or build and to monitor.
Who needs it
Services with large live audiences, viewers spread across many regions, or contractual uptime commitments. A small or regional on-demand service is well served by a single good CDN, and gains more from tuning its bitrate ladder than from adding a second network.
Last reviewed