Прочети на българскиWe have gathered the main benchmark topics in one place to give a clearer picture of what our infrastructure shows in practice. The idea here is not to publish random synthetic numbers, but to show real tests with context, methodology and honest limitations.
What we have measured so far
- Comparing the storage pools in Proxmox: Start, Power, Plus and Micron
- Comparison table: Start vs Power vs Plus vs Micron
- CPU and RAM comparison with sysbench: Start-like versus Power-like
- Network test with iperf3: Start-like versus Power-like
A short summary
- the storage pools are not the same and have to be measured separately
- the CPU difference between a lighter and a stronger profile is clearly visible, mostly under multi-core load
- the RAM tests must be read carefully, and it is good to have more than one memory profile
- the internal network shows strong throughput, but also a real asymmetry between the two directions
What we do NOT claim
- that a single synthetic test on its own describes the whole behaviour of a VPS plan
- that a benchmark of the host is the same as a benchmark inside a real instance
- that a single result is enough for a marketing conclusion
What comes next
- more application-level tests when they make sense: MySQL, PostgreSQL, restore, backup, unpack and update scenarios
- more repeatable tests when needed, if we want to validate a specific profile or an infrastructure change
If you are looking for a VPS profile with a real engineering setup behind it, not just marketing promises, see also the VPSStorages VPS plans.
For more background on the technology see what ZFS storage is and why it matters.