You are here

Home » Blogs » Konstantin Boyandin's blog

Mark your bench

Every user, sooner or later, starts to consider the performance of the hosting plan planned to be bought.

Myself, beyond reading first many reports and reviews (when they are) or outright experimenting (when the hoster is little known), did my own benchmarking.

This is what I am doing now for every VPS I used and use and the conclusion is this: there are no tests, benchmarks, anything measurable in digits, that could 100% correlate with how happy you will be with the hosting plan you'll get.

Poor benchmarks can mean a given node (server etc.) was overloaded at that very moment, and/or about to fail. That's sad, but **it happens. Everything leaks and everything fails, no exceptions.

Good benchmarks could indicate the VPS was sitting alone, oh yes alone on an unused node. Later, when more VPSes sit around and spend the same hardware, real-life parameters can become very, very different.

Regular benchmarking, along with using many other criteria (network efficiency, support quality, good old common sense and intuition, after all) is what can help.

From now on I will do benchmarking of everything I had experience with, and everything I'd like to test, VPSes, cloud servers, you name it. Even though the benchmark isn't the ultima ratio, it can give some thoughts. And yes, benchmark tests will be re-run on everything I use, from time to time.

And, to mark really convenient kind of benches my sites ever sat on, I'll post recommendations. There are many VPS/cloud providers I have good experience with and would recommend them to others. Quality, stability and general impression. Oh yes, and price within reasonable bounds.

And, in case that could help anyone, I'll give longer instructions of using the same tools I use, to those wishing to mark their benches. Stay tuned.