I’ve read quite a few storage press releases talking about how “product X is capable of performance Y and IOPs Z.” I also notice that they didn’t say “we measured this, this way, and this is what we found.” I wonder why.
I look at it this way, if we reported numbers the way lots of these folks report numbers, our JackRabbit JR5 machine would have a bandwidth of 6.2GB/s read and 5GB/s write. Since this isn’t what we measure, these numbers are not relevant. They are a pleasant fiction.
I’ve seen people gleefully reporting the 50-100k IOPs from SSD (disk channel Flash) when applications see 2-10k at best. I still see many reports of 3GB/s for PCIe cards, when what is measured is under 1.5GB/s.
Yeah, this is annoying. I am a fan of good and honest marketing, and I am happy to benchmark our stuff versus pretty much anything out there.
I hope that people are skeptical when they hear these numbers and ask for the test cases and input decks, and access to machines to run them. Performance reality is often very different than what is suggested in these marketing bits.