Upgrading Our Network Transit
More upstream bandwidth, better peering, and what it means for latency on your server.
More upstream bandwidth, better peering, and what it means for latency on your server.
Network capacity is one of those things you only really notice when you don't have enough of it. For most of 2024 our upstream transit was doing fine — but as we added more customers and our DDoS mitigation started handling larger volumetric attacks, the headroom on our transit links started to look uncomfortably thin.
We've spent the past few months renegotiating transit agreements and adding new upstream providers. The result is roughly 3x more available bandwidth than we had at the start of last year, across a more diverse set of upstream ASNs.
Having a single upstream provider is a single point of failure. If that provider experiences congestion or an outage, so do you. By maintaining relationships with multiple transit providers, we can route around problems automatically via BGP. If one upstream is congested or having issues, traffic shifts to the others without any manual intervention.
Better peering with major UK and European networks has had a measurable impact on latency for players connecting from those regions. We've seen average ping reductions of 8–15ms for EU-based players on some routing paths, depending on their ISP. It's not dramatic, but in a game like Minecraft where ticks happen every 50ms, shaving 10ms off your round-trip time does make a tangible difference to feel.
Your server IP, your panel access, your configuration — none of that changes. The network upgrades happen entirely at the infrastructure layer. You just benefit from the result.