Each server includes two redundant links for the public network and two for the private network. Link speeds of 1 Gbps, 10 Gbps, 25 Gbps and 100 Gbps are available depending on location.
Carries traffic to and from the internet, through core routers connected to several Tier 1 carriers.
The Global Private Network connects your servers across every data center with no charge for the traffic.
A separate out-of-band (OOB) network for hardware management, hidden behind a virtual private network (VPN) by default.
Carrier and exchange partners vary by data center, and together they give your traffic many short routes to its destination.
Round-trip times across the network measure how long traffic takes to reach another data center or cloud region and come back, which decides whether an auction, a frame, or a trade lands on time.
| Network | Fastest | Longest |
|---|---|---|
| Public | Under 1 millisecond (ms) between data centers in the same city | 352 ms between Singapore and São Paulo |
| Private | Under 1 ms between data centers in the same city | 380 ms between Hong Kong and São Paulo |
| Google Cloud | Under 1 ms between Google US East4 and Servers.com Washington, DC | 318 ms between Google South America East and Servers.com Singapore |
| Amazon Web Services (AWS) | 1 ms between AWS US East 1 and Servers.com Washington, Virginia | 362 ms between AWS South America East and Servers.com Singapore |
How the network stays up when something fails comes from redundancy in the hardware and in the way traffic is routed.
There are five switches per server. Two for Public Network, two for Private Network and one for OOB access.
Hyperscale-grade network design with active-active redundancy on both public and private networks.
Public core routers connect to multiple Tier 1 carriers, and private network routers keep at least two independent connections.
Our intelligent routing selects the best available path for your internet traffic while maintaining multiple alternative paths for redundancy, which improves both fault tolerance and performance.
The fabric has no switching loops, so a failed switch or Link stays contained and doesn’t spread across the network.
By default, each pair of ports is bonded into one link, so traffic reroutes automatically if one fails. Ports can also be used separately.
Beyond standard border gateway protocol (BGP), our proprietary routing optimization engine continuously measures network performance and automatically steers your traffic around provider congestion and bottlenecks delivering lower latency and consistently fast connectivity to your users worldwide.
Protection against the most common amplification distributed denial-of-service (DDoS) attacks, free in every location.
Native support for IPv4 and IPv6 addresses on the public network.
Combine Layer 2 (L2) Segments, Direct Connect, Firewall, domain name system (DNS), and virtual private network (VPN) to create the topology, routing, connectivity, and security your workloads need, with the control and performance of bare metal infrastructure.
Support from engineers, backed by a team that knows your environment, is part of every product we sell, and it doesn’t depend on how much you spend with us.
Engineers answer your messages 24 hours a day and every day of the year, and your account team brings the context of your environment.
Your account managers stay with you for the life of the relationship, from the first call onward.
Architects take part in design, migration, and later changes, so decisions get made with the workload in view.
We work in your Slack, Discord, Teams, Telegram, or WhatsApp instead of asking you to learn a ticket system.
Yes, when it runs over the Global Private Network, which connects all our data centers.
Up to 20 gigabits per second (Gbps) for each dedicated server.
Our routed fabric gives each server its own isolated segment, which is more resilient. L2 Segments add shared segments where your software needs them.
Carriers vary by data center, and the list for each location is on our available internet protocol (IP) carriers page.
Check round-trip times for every location, or ask an engineer to measure the paths that matter to your platform.