Every server is connected to a network we operate ourselves in each of the six locations, with two 100 Gbps uplinks per site, always-on DDoS protection and native IPv6. This page covers addressing, reverse DNS, what happens during an attack, traffic accounting, port speeds and the tools you can use to measure us before you order.
IPv4 and IPv6
Each server receives one dedicated IPv4 address and a routed /64 IPv6 block. Both are configured automatically on our images; the dashboard shows the addresses, the gateway and the prefix for manual configuration on custom installs. Addresses are not shared with other customers and stay with the server until it is deleted, including across reinstalls and resizes.
Dedicated servers can request additional IPv4 addresses (up to a /29 per server, with an RIR justification) and an IPv6 /56. Customer-owned prefixes can be announced over BGP from dedicated servers: open a ticket with the prefix, the ASN and a letter of authorisation, and we set up a session with prefix filters on our side.
Reverse DNS
Set the PTR record of any IPv4 address, and of any address in your IPv6 block, from the server page in the dashboard. Records propagate within minutes. Mail servers should set the PTR to the hostname that the server presents in its HELO, and that hostname should resolve back to the same address:
dig -x 203.0.113.10 +short
dig A mail.example.com +shortDDoS protection
Volumetric attacks are detected within seconds by flow analysis at the edge. Traffic towards the targeted address is diverted into scrubbing, where attack packets are dropped and clean traffic is forwarded, so legitimate users keep reaching the server. The filtering capacity is 3 Tbps and is included with every plan at no extra cost. You receive an e-mail when mitigation starts and ends, and the event appears in the dashboard with the peak rate and the vectors seen.
Mitigation is transparent for TCP services. UDP-based services (game servers, VoIP, DNS) are protected with protocol-aware rules; if a legitimate UDP protocol is affected, tell us the ports in a ticket and we adjust the profile for that address. Application-layer attacks (HTTP floods) are not covered — a reverse proxy or a CDN in front of the server handles those.
Traffic allowance
| Product | Port | Monthly allowance | Beyond the allowance |
|---|---|---|---|
| VPS, RDP | 1 Gbps | 4 TB to 50 TB depending on the plan | Throttled to 100 Mbps, never billed |
| Dedicated Ryzen, Storage | 1 Gbps | Unmetered | — |
| Dedicated EPYC, Xeon | 10 Gbps | 100 TB or 200 TB | Limited to 1 Gbps, never billed |
| GPU | 1 to 25 Gbps | Unmetered | — |
Only outbound traffic counts towards an allowance. Traffic on the private network between your servers in the same location is free and not counted. Usage is shown per server in the dashboard and through the API, and resets on the renewal date.
Port speeds
Ports are dedicated to your server and never oversubscribed at the rack level: a 1 Gbps port delivers 1 Gbps in both directions at the same time. VPS, RDP, Ryzen and Storage servers use 1 Gbps; EPYC and Xeon dedicated servers use 10 Gbps; GPU nodes use 1, 10 or 25 Gbps depending on the plan. Each location connects to the internet through two 100 Gbps uplinks and to major exchanges — DE-CIX in Frankfurt and New York, AMS-IX in Amsterdam, FICIX in Helsinki, Any2 in Los Angeles and SGIX in Singapore — plus several transit carriers.
Looking glass and speed test
Before ordering, measure us. The looking glass opens TCP connections from our edge to any public host you name and reports the latency, so you can see how far your users are from each site. The speed test page lists download files and an iperf3 endpoint per site:
curl -o /dev/null https://cheapserv.com/dl/100mb.bin
iperf3 -c fra1.speed.cheapserv.com -p 5201 -P 4
iperf3 -c sin1.speed.cheapserv.com -p 5201 -REndpoints follow the pattern <site>.speed.cheapserv.com for fra1, ams1, hel1, nyc1, lax1 and sin1. Use -R to measure the download direction and -u for UDP.
Acceptable traffic
VPN servers and proxies are welcome on every product. Tor relays are allowed; Tor exit nodes are not, because of the abuse load they place on the address space. Cryptocurrency mining is allowed on dedicated and GPU servers only. Port scanning of networks you do not control, spoofed traffic and attack traffic of any kind lead to immediate suspension under the acceptable use policy.
Private network
All your servers in one location share a private VLAN on a second interface — 1 Gbps on VPS and RDP, 10 Gbps on 10 Gbps dedicated servers, up to 25 Gbps on GPU nodes. It is unmetered and isolated from other customers. Addressing is yours to choose; use it for replication, backups, storage and cluster traffic so that nothing sensitive crosses the public interface.
Something missing or wrong on this page? Open a ticket and tell us.