"Brown, Mike" <Mike.Brown@CoreExpress.net> writes:
> I'm intrigued by your comment that "... it's been statistically proven that
> sampled netflow
> is perfectly adequate for determining traffic patterns as well as accounting
> usage."
>
> Can you provide any references or other information to the analyses of the
> accuracy of sampling for accounting?
I did some studies myself that were never put into a formal report.
What I found was that if you want 95th percentile numbers, anything
over 1/3 yields lower bills. For traffic engineering, much lower
sampling rates are practical.
The one empirical data point that I can think of off-hand is that if
you sample at 1/5, the 12 hour average is the same (+/- 0.5%) as the
12 hour average as reported by interface counters. Even at 1/5, as
the time window decreases, the error rate increases significantly.
-- Matt Ranney - mjr@ranney.com
This archive was generated by hypermail 2b29 : Mon Aug 05 2002 - 10:42:42 EDT