Re: IP SLA history - please explain

Disclaimer: Did you ever study statistics? I didn't :-)

Those commands relate to the mathematical science of statistics, though.

I think if these commands were entered in the configuration of an ICMP
echo IP SLA operation, you'd end up with 10 groups ("distributions",
"history distributions-of-statistics-kept") of statistics broken up
into 100ms ("history statistics-distribution-interval") chunks:

distribution 1: ping RTTs from 0-99ms
distribution 2: ping RTTs from 100-199ms
distribution 3: ping RTTs from 200-299ms
distribution 4: ping RTTs from 300-399ms
distribution 5: ping RTTs from 400-499ms
distribution 6: ping RTTs from 500-599ms
distribution 7: ping RTTs from 600-699ms
distribution 8: ping RTTs from 700-799ms
distribution 9: ping RTTs from 800-899ms
distribution 10: ping RTTs from 900-infinity ms

I _think_ this would allow you to do things like say "within the
400-499ms range, the minimum was 405ms and the maximum was 489ms. the
average over the 173 ping RTTs within this range was 438ms". This
would only be useful if you were performing some very granular
analysis of network performance -- by default, there is one group
("distribution") covering all measurements (all values from 0ms to
infinity).

I'll leave it to someone with more clue in this area to provide better
practical examples of usage for the above, and for the 'lives-kept'
and 'buckets-kept' commands.

I'm sure, given enough time with the IP SLA config guide and command
reference, it would all become clear(er).

cheers,
Dale

2009/5/4 nowy1981 <freemaxis_at_gmail.com>:
> Could you explain exactly what the following values exactly do?
>
> history distributions-of-statistics-kept 10
> history statistics-distribution-interval 100
> history lives-kept 2
> history buckets-kept 60
>
> Thanks

Blogs and organic groups at http://www.ccie.net
Received on Mon May 04 2009 - 12:09:58 ART