FCC outage reports

Sat Sep 19 21:30:30 UTC 1998

shields at crosslink.NET (Michael Shields) writes:
>You'd also want to look at more than just the count of outage reports;
>the size of the outages matters, the total size of the network, and
>the MTTR.  If you "had to" distill it to a single number I suppose
>percentage uptime is the one to use.

The problem with percentage uptime is the law of large numbers.
You need a huge disruption affecting a lot of customers for a long
time to move the uptime percentage.  As a customer I don't really
care how much everyone else circuits were up.  I only care about
how much my circuits are up.  If my AT&T circuit is down, it doesn't
matter to me that 99.99999999999999999% of all AT&T's other circuits
are working.

>I'm not sure fiber cuts are common enough to make a meaningful
>statement about the reliability of one carrier vs. another.

I agree.  The initial fiber cut isn't that meaningful.  I have no
hypothesis why backhoes (or other natural hazards) would prefer
one carrier's fiber over another carrier's fiber.  WilTel's initial
plan of putting fiber inside steel gas pipelines not withstanding.

Hence my original statement that there is no difference between the
carriers based on just fiber cuts.  Its like calculating the probability
of a coin flip, just because one carrier had fewer fiber cuts this year
doesn't seem to be a good predictor how many fiber cuts they will have
next year.

But how the carrier reacts after the fiber cut is an area you can
make statements about the differences between one carrier vs. another.
-- 
Sean Donelan, Data Research Associates, Inc, St. Louis, MO
  Affiliation given for identification not representation