[PLEU-BEAST]
40 GbpsDedicated
- Processor
- 2 × Gold 6152
- Memory
- 128 GB DDR4
- Storage
- 1 TB NVMe
- Est. setup
- 2–3 days
€2,699 / mo
OrderBurst headroom and high-volume delivery
40 Gbps Servers
40 Gbps dedicated servers
Size an origin for the hour when several CDN edges request fresh content together. Choose the actual port listed on the plan, then check whether your disks can sustain those reads.
Workload illustration, not live server measurements.
- Ports listed
- 25 Gbps / 40 Gbps / 50 Gbps
- Offshore networks
- 3
- Plan count
- 4 configs
- Best for
- Origins / clusters
- Review
- Before invoice
- From
- €2,199 /mo
Server configurations
40 Gbps Servers
4 packages EUR monthly
The port is the server connection speed. It does not specify a minimum sustained rate, a monthly transfer allowance or an overage price. Ask for these terms for the exact package before ordering.
Offshore Network I
2 packages[PLEU-BEASTX]
50 GbpsDedicated
- Hardware
- Configured to order
- Est. setup
- 2–3 days
On quote
Request quoteOffshore Network II
1 package[NLEU-APEX]
25 GbpsDedicated
- Processor
- 2 × E5-2673 v4
- Memory
- 128 GB DDR4
- Storage
- 1 TB SSD
- Est. setup
- 2–3 days
€2,299 / mo
OrderOffshore Network III
1 package[CHEU-ULTRA]
25 GbpsDedicated
- Processor
- 2 × AMD EPYC 7452
- Memory
- 128 GB DDR4
- Storage
- 1 TB NVMe
- Est. setup
- 2–3 days
€2,199 / mo
OrderHigh-volume architecture
A 40 Gbps server is usually part of an architecture: a CDN origin, a large relay, a transcode-output node, or a high-volume cache. It gives space for traffic bursts while the rest of the system catches up.
It should still be sized honestly. Disk reads, codec choices, cache misses, middleware behavior, viewer ISP paths, and failover design can all become bottlenecks before the port itself.
Use cases
Where 40 Gbps earns its keep.
A cold cache can change the bottleneck: 30 Gbps of origin reads is 3.75 GB/s of payload before overhead. Check storage read speed with your real file sizes as well as the route to your edge networks.
Large VOD origin
Workload: A high-read origin serving edge nodes or viewers directly.
Sizing: Disk read patterns, hot-title placement, and cache strategy affect real throughput.
Fit: 40 Gbps gives a larger safety margin for bursty catalog demand.
Limit: If a single node becomes operationally risky, split origin and cache roles.
IPTV cluster node
Workload: Live-channel delivery with multiple downstream relays or load-balanced frontends.
Sizing: The platform should model stream bitrate, concurrency, and failover capacity.
Fit: 40 Gbps is a strong cluster building block before moving to 100 Gbps.
Limit: Do not size on viewer count alone; codec, bitrate, and middleware overhead matter.
Transcode farm egress
Workload: Workers that transcode media and push outputs to storage, viewers, or cache.
Sizing: CPU/GPU throughput must match the outbound network target.
Fit: 40 Gbps keeps the port from bottlenecking multi-worker output.
Limit: Compute limits often arrive before network limits in transcode-heavy systems.
Calculate cache-fill bandwidth from origin logs
Provisioning
Choosing an operating system
For a cache node, confirm the NIC driver, storage controller and filesystem support. Test cold-cache reads before moving traffic.
Check installation, driver and licensing requirements
Debian stable
Ubuntu Server LTS
AlmaLinux
Rocky Linux
Fedora Server
Windows ServerQuestions & answers
40 Gbps Servers: buying questions
Browse all topicsWhy are 25 and 50 Gbps options listed with 40 Gbps servers?
They are neighbouring choices for the same buying decision. Use the dedicated port printed on each package and request a quote for a custom build. The page title does not change the port you order.
Why can a video origin stutter below 40 Gbps average usage?
Short bursts can fill an interface queue even when the longer average looks low. Compare short-interval egress, drops and disk latency during player retries rather than using monthly traffic to diagnose a burst.
What read rate does a 40 Gbps origin need?
The port represents 5 GB/s before overhead. Cold cache fills can demand sustained disk reads close to the application delivery rate; warm content can come from memory. Test with the files and cache state you expect.
Can I move cache refreshes outside a live event?
Yes, if your application controls refresh timing. Stagger cache fills and pre-position popular assets so origin fetches do not compete with live delivery at the start of the event.
What should I supply for a custom high-bandwidth quote?
Include sustained and peak egress, burst duration, hot working-set size, required disk redundancy and recovery targets. Identify the audience networks and the package you are using as a starting point.
Related paths
Explore related infrastructure.
Compare adjacent bandwidth, media, and pooled-capacity routes before choosing a configuration.