HostUp: High Availability VPS is now standard option where data is replicated to 3 different servers

HostUp now offers VPS with Block Storage featuring High Availability as standard for all their VPS. This is the default option and comes at no additional cost. What this means is that instead of traditional RAID where data is replicated to disks on the same server, the storage is now replicated to 3 separate servers.

Even in the event of complete server failures the data is still available and VPS can reboot within a few minutes on another server. In terms of redundancy this new offering marks a significant upgrade from traditional RAID, taking data replication and system resilience to unprecedented levels. But it's not all about redundancy, it also got a lot higher throughput!

Storage over network is not slow

For most servers local storage is used. Typically if one would want to combine redundancy with performance RAID-10 is used. With 4x NVMe disks this would result in some pretty impressive benchmarks inside the VPS. However we all know that more disks means more throughput.

Adding more than 4 disks per physical server seldom makes sense, however now with each VPS having a network-based storage, powered by Ceph the throughput that can be achieved is much greater. This is because each VPS now can read and write to the entire cluster consisting of 30+ individual NVMe.

We've ran a quick fio benchmark on a test VM to demonstrate this:

And with a larger block sizes: