[smokeping-users] Packet loss calculations in Smokeping

[Tobias Oetiker]

Hi Vern,

Today Vern.Dias at VerizonWireless.com wrote:

> There are never any NAN values that I have found in the actual latency
> fields, regardless of the number of lost pings.  To me this means that
> any attempt at having rrdtool calculate an accurate packet loss value is
> doomed from the start. Looks to me like rrdtool is attempting to
> calculate all latency values.

note that the first 3 fields in the RRD file are not round trips,
but the results of the analysis of the pings sent out in each
round. We store this data explicitly since it can not be rebuilt
later from what is stored in the rrd file:

0 - uptime - this is for measuring ip retention time with dynamic ips
             this is set to NaN unless you use this feature

1 - loss   - the number of lost packets

2 - median - the median round trip time

3-n - ping[1-n] - the ping latency

the roundtrip times are 'sorted'. if not all pings are successful
the successfull ones are positioned n the middle

> Seems to me, that in the case of a packet loss situation within a single
> poll, it is up to Smokeping to feed actual data from all 10 polls to
> rrdtool, even if one or more or all of the polls are timeouts.

it does this ... but as you explained earlier, since the pings do
not arrive at 'step-time' it takes normally two losses in a row to
manifest them selfs in the log ...

regarding your 'analysis' ... don't you think something so fatally
flawed would have triggerd some discussion in all the years
smokeping has been around ? if you check the mailinglist archives
you will see that it has and eventually people understood how the
data is handled, and why it happens the way it does ...


cheers
tobi

-- 
Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten
http://it.oetiker.ch tobi at oetiker.ch ++41 62 213 9902