|
Hi Luca, I didn't realize that RRDTOOL could be used to perform curve-fitting. Thank you for the pointer to this link. From this link, I infer that the dark blue 'Trend' line would represent an attempt to perform a non-linear curve fit using a sliding window average across 30 minutes. [I don't see a dark blue line on this chart, though, for reasons I don't understand.] More interesting to me -- I'm still struggling to understand Upper and Lower. Perhaps these are two non-linear curve fits as well, using some sort of sliding window average ... but what are they attempting to predict? Perhaps 'worse-case' throughput and 'best-case' throughput? I'm poking through ../report.c ... but losing my way. ? --sk On 4/10/2013 4:24 AM, Luca Deri wrote:
|
_______________________________________________ Ntop mailing list [email protected] http://listgateway.unipi.it/mailman/listinfo/ntop
