Video CDN
A video CDN is a content delivery network set up for streaming video. It keeps copies of video segments and manifests on edge servers, called points of presence, close to viewers, so each request travels a short distance instead of going back to the origin server. Video traffic is heavy and continuous: a player asks for a new segment every few seconds for the whole view, and a live event sends its entire audience to the same segments at once. A video CDN is tuned for that pattern.
How does a video CDN work?
Encoded segments and manifests sit on an origin, the storage or server that holds the master copy. When a player requests a segment, DNS or anycast routing sends it to a nearby edge server. If the edge already has the segment, that is a cache hit and it answers at once. If not, it is a cache miss: the edge fetches the segment from the origin, or from a mid-tier shield cache that protects the origin, stores it and serves it.
On-demand segments never change, so edges can keep them for a long time. Live is different. A live manifest updates every few seconds as new segments appear, so it is cached only briefly, while each new segment is pulled to the edges as soon as the first viewer asks for it. Popular titles end up cached almost everywhere, while the long tail of a large library is fetched from the origin more often.
Video CDN vs. a general-purpose CDN
The machinery is the same: edge caches, routing and an origin. What changes is the workload. A web CDN mostly serves small files such as images and scripts that load once per page. A video CDN serves a steady stream of multi-megabyte segments for the length of every view, often for hours.
That shifts what matters. Cache capacity has to hold large libraries, not just popular pages. Request collapsing matters for live: when thousands of viewers ask for a brand new segment in the same second, the edge should send one request to the origin, not thousands. Throughput to the viewer's internet provider matters more than time to first byte.
One point is often missed. Putting a single MP4 behind any CDN makes it arrive sooner, but it does not make it adaptive. A phone on a weak signal still has to pull the same full-bitrate file. A video CDN pays off when it delivers an adaptive ladder.
What affects video CDN performance?
Distance is the obvious factor, but peering usually matters more. Peering is how directly a CDN connects to each internet provider's network. A CDN can have a server in a city and still route one mobile carrier's traffic through a congested exchange, so performance is measured per ASN, the network identifier of each internet provider, not just per country.
Cache hit ratio is the next factor. Every miss adds an origin round trip, which shows up as slower startup and, on live streams, as stalls at the live edge. Segment length affects how many requests a view makes and how quickly new live segments reach the edges. Protocol support counts too: HTTP/2 and HTTP/3 reduce connection setup on mobile networks.
The reliable way to see all this is from the player. CDN logs show what the server sent. Player-side request latency and throughput, broken down by CDN, region and ASN, show what viewers actually got.
How is video CDN delivery priced?
Most CDNs bill per gigabyte delivered, with rates that vary by region and fall at higher monthly volumes. Some add a charge per request, and cloud providers bill separately for data leaving the origin to fill the edge caches. Larger contracts often bill on committed bandwidth instead, measured at the 95th percentile: throughput is sampled every few minutes across the month, the busiest samples are discarded, and the highest remaining value sets the bill. Video platforms often bundle delivery and bill per minute watched.
Each model moves differently. Per-gigabyte delivery grows with bitrate as well as watch time, so moving viewers to a higher rendition raises the bill even when viewing is flat. Percentile billing forgives a short spike but not a long peak. Ramadan viewing in Saudi Arabia and the UAE gathers after iftar, roughly 7 pm to 2 am, every night for a month, which is the kind of sustained peak that sets a percentile bill. Per-minute delivery tracks watch time and leaves bitrate to the encoding ladder.
How does FastPix handle video CDN delivery?
FastPix delivery is built in, not a separate service to wire up. Every video is packaged just in time as CMAF and served as HLS and DASH across a multi-CDN network of 300+ points of presence. Edge routing is chosen per viewer, and if one CDN degrades in a region, traffic fails over to another. Live streams use the same network, with concurrency that scales with your plan.
Delivery covers more than segments: video, posters, storyboards and images can all be served from your own custom domain. thumbnail images come from the playback ID: pick the frame with time, resize with width and height, choose jpg, png or webp, and set a fit mode such as crop, pad or intellicrop, a smart crop. A webp poster sized to the player keeps the first request on the page small. Images inherit the playback ID's access policy.
Behind the CDN, each upload gets a per-title ladder in H.264 and HEVC (H.265) as standard, with AV1 available on request, so the edge serves renditions a phone can play rather than one heavy file. Playback IDs can be public or signed with expiring tokens.
Video Data shows delivery from the viewer's side, with request latency and throughput broken down by country, region, CDN and ASN, and alerts to PagerDuty, Slack or Opsgenie.
There is no separate CDN bill. Ingress, egress, request charges and cross-region transfer are bundled into the per-minute streaming price, with volume discounts applied automatically. Capping the top resolution at 720p, 1080p, 2K or 4K limits the heaviest renditions. The free plan includes 100,000 streaming minutes a month.
By the numbers
- Netflix lists data use of up to 3 GB per hour for high definition streaming and up to 7 GB per hour for 4K. (Netflix Help: How to control how much data Netflix uses)
