Video-on-demand origins built to binge.
Store huge catalogs on NVMe and HDD, then serve them to thousands of concurrent viewers over unmetered bandwidth. Pair high-density storage with up to 200 Gbps of egress and direct peering to the eyeball networks your audience actually uses.
Catalogs that never stutter.
Dense storage backed by an NVMe cache tier and an unmetered, directly-peered uplink — the three things a VOD origin actually needs.
High-density storage
Up to 8 bays and 96 TB raw per node — keep your whole library on a single origin instead of stitching shards.
NVMe cache tier
Serve hot titles from NVMe while cold catalog lives on high-capacity HDD — best cost per streamed gigabyte.
Byte-range & ABR ready
Efficient range requests and HLS/DASH segment serving for smooth seeking and adaptive bitrate playback.
Peered to the eyeballs
Direct peering with Vodafone, Comcast, Telefónica, BT, Digi and 3,600+ more puts catalog one hop from viewers.
Engineered for the long tail.
The specifics that make a dense catalogue cheap to store and fast to stream.
96 TB per node
Up to 8 bays and 96 TB raw on a single origin — keep the entire library in one place. No sharding and no stitching catalogues across boxes.
NVMe cache, HDD depth
Hot titles serve from an NVMe tier while the cold catalogue rests on high-capacity HDD. It is the cheapest cost per streamed gigabyte we can build.
Range & segment serving
Efficient byte-range responses and HLS/DASH segment delivery come out of the box. Seeking stays smooth and adaptive bitrate switches cleanly.
Unmetered 1–200 Gbps
Pick a guaranteed egress tier from 1 to 200 Gbps, all of it unmetered. A title that suddenly trends never turns into a surprise invoice.
One hop from viewers
Direct peering with 3,600+ eyeball networks puts the catalogue a single hop from most audiences. AI routing steers around congested paths.
Drop-in CDN origin
Point your own CDN edge at the box and let it pull segments on demand. It behaves as a clean origin for any player or cache layer.
Serve the long tail at any volume.
Match the bandwidth tier to your concurrent sessions — every tier is guaranteed and unmetered.
Niche & regional
Indie catalogs, course libraries and regional services with steady on-demand traffic.
Growing platforms
Busy VOD platforms serving large catalogs to tens of thousands of concurrent viewers.
CDN origin
Origin behind your own CDN edge, pushing sustained tens of gigabits to cache nodes.
Serve every codec & container.
It is your origin and your packager — we just move the bytes. Common formats and what they buy you.
| Codec | Containers | Adaptive | Best for |
|---|---|---|---|
H.264 / AVC universal | MP4, fMP4, TS | HLS / DASH | Widest device support |
H.265 / HEVC ~40% smaller | fMP4 / CMAF | HLS / DASH | 4K at lower bitrate |
AV1 next-gen | MP4, WebM | DASH | Best efficiency, modern players |
VP9 royalty-free | WebM | DASH | Web & open pipelines |
Packaging CMAF low-latency | HLS, MPEG-DASH | ABR ladders | One package, every player |
Pick your VOD origin.
Storage origin
- 8× 12 TB — 96 TB raw
- Unmetered uplink for cold catalog
- Ideal whole-library single node
NVMe cache origin
- Single EPYC 9254 · 10× NVMe
- Hot-title cache + unmetered uplink
- Online within the hour
100 Gbps hybrid
- Dual EPYC 9554 · 512 GB
- NVMe + HDD, 100 Gbps unmetered
- Behind your own CDN edge
From upload to first play.
Pick a region
Deploy the origin in the PoP closest to your biggest audience — 40 to choose from across Europe and the US.
Load your catalog
Sync your library over the unmetered uplink and lay out your NVMe cache tier for the hot titles.
Point your CDN
Set the server as origin for your CDN or player. Range requests and ABR segments stream straight from disk.
Packagers & players.
Run the media stack you already know — a few that work well on a ZetServers origin.
Standard on every build.
NVMe + HDD layout
Every origin ships with an NVMe cache tier in front of high-capacity HDD, laid out for your hot-versus-cold split.
Unmetered to 200 Gbps
A guaranteed, unmetered uplink scaling from 1 Gbps to 200 Gbps — matched to your concurrency, never billed by the gigabyte.
RAID & geo-replication
Run ZFS or hardware RAID across the bays, with optional replication to a matched node in a second PoP for geo-redundancy.
IPMI & full root
Out-of-band IPMI plus full root from day one — reinstall, remount and re-tier the storage entirely on your own schedule.
Bucharestro
Dublinie
Milanoit
Viennaat
Frankfurtde
Amsterdamnl
Sofiabg
Madrides
Lisbonpt
ParisfrMarseillefr
Warsawpl
Istanbultr
AshburnusVideo-on-demand questions.
How big a catalog can one node hold?
Storage builds reach 8 bays — up to 96 TB raw with 12 TB drives. Add an NVMe tier for hot titles and keep the long tail on HDD.
Can I use it as a CDN origin?
Yes. Many customers run these as the origin behind their own CDN edge, pulling segments over a 25–200 Gbps unmetered uplink.
Does it handle byte-range requests?
Fully — efficient range serving and HLS/DASH segment delivery mean smooth seeking and adaptive bitrate for every viewer.
What about redundancy?
Run ZFS or hardware RAID across the bays, and replicate across two PoPs for geo-redundancy. We can spec matched nodes.
Is egress metered?
No. Pick a tier from 1 to 200 Gbps — it is guaranteed and unmetered, so a viral title never produces a surprise bill.
How fast is delivery?
Direct peering with 3,600+ networks keeps your catalog one hop from most viewers, and AI routing avoids congested paths.
Which media servers can I run?
Anything you like — you have full root. Teams run nginx or Varnish as the origin, Jellyfin, Emby or Plex for library apps, and custom packagers for HLS and DASH.
Can I pair storage with a separate egress node?
Yes. A common layout keeps the deep catalogue on a storage origin and fronts it with a separate NVMe cache or edge node, linked over our backbone.
Scaling a catalog? We'll spec the origin.
Tell us your library size, expected concurrency and CDN setup — we will design the right storage-plus-egress origin, with a price match on comparable quotes.