CDN Distribution
Every encoded asset is synced to the edge through our CDN dashboard. You can track local, synced, partial and failed state per title, force a sync, or invalidate cache for an entire library in one click.
Sync status
Local, synced, partial and failed states per asset.
One-click ops
Sync All and Invalidate All Cache actions.
Size awareness
Live totals of files synced and storage used.
Selective invalidation
Purge a single title or whole catalogue.
Origin, edge and cache behaviour
Encoded segments are written to origin storage and pulled to the edge on first request, then served from cache for everyone after. Because segments are immutable once published, they can be cached for very long periods; manifests are the part that must stay fresh, especially for live and DVR playback where the playlist changes every few seconds.
Getting those two cache policies right is most of CDN tuning. Long-lived segments with a short-lived manifest give high offload and correct behaviour; a single blanket TTL either breaks live playback or leaves the origin serving traffic it should never see.
Keeping the catalogue in sync
The CMS tracks per-asset delivery state - local, synced, partial or failed - so operations can see at a glance which titles are actually reachable from the edge and which are still only on origin. Partial state usually means an interrupted upload or a rendition that failed to encode, and it is the failure mode most likely to reach viewers as a title that plays at one quality but not another.
Bulk actions cover the routine work: sync the whole catalogue after a migration, or invalidate cache for a single title after re-encoding it. Invalidation is deliberately scoped, because purging an entire distribution forces every segment to be re-fetched from origin and produces an expensive traffic spike.
Cost, monitoring and multi-CDN
Delivery is usually the largest recurring cost in a streaming service, and it is driven by three things: the top rungs of the bitrate ladder, cache hit ratio, and how much traffic is served from the wrong region. Monitoring offload per region alongside player-side quality metrics shows which of the three is actually costing money.
Multi-CDN - splitting traffic across two providers and steering by measured performance - is worth the added complexity once traffic is large enough or availability requirements are strict. Below that scale a single well-configured CDN with a tested origin failover is usually the better trade.
Inside the CMS
Frequently asked questions
- Which CDN do you work with?
- The pipeline is CDN-agnostic. We integrate with the provider you already have a contract with, or recommend one based on your audience geography and expected peak traffic.
- How fast does a cache invalidation take effect?
- Typically seconds to a few minutes depending on the provider and how wide the purge is. Because of that delay, a replaced asset is usually published under a new path rather than purged in place.
- Can we serve from our own origin?
- Yes. Origin can be object storage you own, our storage, or a hybrid where the catalogue lives with you and only the delivery layer is managed.