I have juste discovered a study published on Rasmus’s Lerdorf blog, it is titled Low-Cost VPS Testing. The tests are not driven by scientific standards as it is written by author. But it is not really necessary because speed test are always empirical in the hosting domain. Many factors can restrict power and speed (most of the time, shared usage of bandwidth, cpu, RAM or hard drive usage are explaining speed drops) and it is reasonable to limit tests to quick tests, done with common tools. This why he has briefly tested CPU speed by compiling a common PHP version, the hard drive bandwidth by using dd command for file synchronisation and the network bandwidth with iperf 3 speed test.
Considering that the same tests have been strictly ran on all the VPS offers he has tested, we can look at results with confidence. And to use them to choose the best offer we want regarding our own criteria.

The tested providers are mostly USA based, this is why I don’t know a part of them. So I have looked at the providers I already knew or that I have already used. My web projects are hosted by Amazon Web Services on Lightsail VPS and I work with one of my customers on a website hosted on an OVH VPS. Some time ago, I have worked with some DigitalOcean VPS. I am happy to discover that Amazon Lightsail solutions have a pretty satisfying average performances. Not the bests but strong and indicating that Amazon provides to me serious solutions with dedicated hardware ressources which seems not to be the case of some others providers presents in the test.

Conclusion: I keep my Amazon Lightsail VPS and I let you discover other results on Rasmus Erdorf blog 😉