Failure Rate for Heimdal SW?

["Henry B. Hotz" <hotz@jpl.nasa.gov>]

I have one of those availability requirements that says Kerberos can't  
fail for more than X hours per year.  Now I've got good historical data  
on the hardware/OS that I can put in that form.

Anybody got any "software failures" which are remotely statistical in  
nature?  Any ideas on how to assign an availability number to the  
software component?  Any failures that require/cause a Kerberos  
restart?
------------------------------------------------------------------------ 
----
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz@jpl.nasa.gov, or hbhotz@oxy.edu