No, it really is downtime. When you go to an openID site, you shell into your server, run this program, authenticate at the site, then stop the service. 99.999% downtime.
If not: I thought OpenID was supposed to make things easier! How does ssh'ing into my server and starting and stopping services make anything easier!?
And why stop the service, anyway? If it's not being used, shouldn't consume much resources. Unless it's hopelessly insecure or something, in which case it should be abandoned anyway.
This seems very cumbersome. I have a hard time believing anyone is more efficient at this 7 step process than using a "normal" OpenID provider. Enter a password, click a couple buttons, you're done.
Why not just run any one of dozens of existing OpenID provider packages? If you really want the 99.999% downtime "feature", just start and stop the server when you want to log, in the same way.
