I think there is a fundamental misunderstanding caused by the OP's article. It mischaracterizes your blog post as a performance test of microdot, but as you say, it has nothing to do with it. Still, a baseline, best-case RPS number is a useful number to associate with microdot for various reasons; it cannot, of course, take the place of testing the real application, but what it can do is give implementers and upper-bound on IOPs to a device via HTTP/S. Take that as you will. Microdot is a very cool project, kudos for doing it and sharing it.