Where your server actually runs
We don't run on someone else's cloud markup. RavenousMC servers run on our own hardware, in a single carrier-neutral facility in the Intermountain West.
Geographically central for North America — low single-digit millisecond latency to most western US ISPs and sub-60 ms coast-to-coast.
Carrier-neutral, SSAE-18 facility. Redundant power (A+B feeds, UPS-backed, on-site generators), N+1 cooling, 24/7 staff on-site.
Two independent 10 Gbit links into the facility's blended transit. If one upstream provider blips, the other carries the traffic — no single carrier outage takes us offline.
Hardware
We run on enterprise hardware that's a generation old enough to be proven and a generation new enough to still ship security firmware. No consumer kit, no rented bare-metal you can't see the model number of.
PowerEdge R-series — iDRAC9 lights-out management, redundant PSUs, hot-swap drives. Standard datacenter rack mounts, no custom whitebox surprises.
Server-grade silicon — high base clocks (matters for Paper's single-threaded main loop) plus enough core count to amortize plugin and chunk-gen work across dozens of concurrent customers.
Standard plans run on enterprise SSD pools; Premium plans run on NVMe — both with regular world snapshots and restic-encrypted off-server backups. See pricing for the per-plan breakdown.
Why it matters
Minecraft servers are sensitive to two things most hosting providers gloss over: single-thread CPU performance (because the main game tick runs on one thread) and disk IOPS (because chunk loads and entity ticks hammer the world directory on every action). Picking a real datacenter with enterprise hardware means your TPS doesn't crater when somebody else on the host decides to run a benchmark.