>>>>OK, but all you're really measuring is the benefit of "SSD" (whatever that is) vs "regular" hosting at one host, on a certain date/time, with a certain (unknown) load level on the host's back-end infrastructure. You would probably get different results at the same host on a different date/time. And there's no way you could extrapolate your results to what you might expect at a different host. Really all you can say is that "SSD hosting" is faster than "regular" hosting - which you'd expect, since your host charges more for it ;)
>>>
>>>I've had the chance to run the same process (not quite identical, few lines were different, but not decisively) several times on the same server, on different times of day. The time it took varied between 3 and 5 hours (it converts about a gigabyte of data from one SQL db to another). The server is a hosted VM somewhere in, I guess, UK.
>>
>>Do you know if that's on a public cloud, or a private/corporate one?
>
>I know the guys are renting it for a number of years, even before cloud became an IT buzzword. "Co-hosted" or something was the buzzword then.

Maybe colocated? That would mean their own server(s), but on the premises of a 3rd party.

If that's the case I'd expect your results to be more deterministic, unless you know it's busier with other stuff at certain times of the day.