[rrd-users] Sup-Sample Aggregate Functions?

[Tim Soderstrom]

Hi All!

I was curious as to if it was possible to create a graph that contains a 
sub-sample aggregate. What I am trying to do is basically smooth out a 
series of points to produce a sort of trend. This actually works well 
but the problem is that I am using a step-size of 2 seconds and am only 
recording data for a short period of time (a few minutes). As a result, 
the graphs look very blocky since you can see jaggedness. Here are some 
examples:

http://www.moocowproductions.org/temp/questions.png

You can see the Trend (using HWPREDICT here) and the Average (using the 
TREND VDEF) come close, but I really want a spline-based smoothing or 
something along those lines.

Something, for instance, similar to this:

http://graphpite.sourceforge.net/samples/index.php?sample=03

Really, I don't care about the actual datapoints. I care about making a 
smooth looking graph that is an estimate of the actual datapoints. I 
checked #rrdtool and was told it was not possible with the standard 
version of rrdtool. Anyone have any ideas on how I could augment rrdtool 
for this or had another suggestion on what I could use?

Thanks!

Tim S.