|
Correlation Table for Active and Passive MonitoringConnie Logg, Les Cottrell and I-Heng MeiIEPM-BW results | Bulk throughput measurements | Bulk throughput simulation | Windows vs. streams | Effect of load on RTT and loss | Bulk file transfer measurements | QBSS measurements | SC2001 Bandwidth Challenge |
At the same time we are capturing the passive NetFlow records and we identify which flows go with which active application call. We add up the throughputs for all the streams for a given application call. We calculate the passive throughput using three methods:
R is the
correlation coefficient.
The Correlation Coefficient R is defined as:
R = COV(x,y) / (Std(x) * Std(y)),
where:
COV(x,y) = (1/(n-1)) * SUM((xi-Avg(x))*(yi-Avg(y))),
is the
covariance,
and the SUM is over the n paired Samples (i=1..n) , also
Avg(x) = (1/n)) * SUM(xi)) and
Avg(y) = (1/n) * SUM(yi)) are the averages, and
Std(x) = sqrt((1/(n-1)) * SUM((xi-xm)2)) and
Std(y) = sqrt((1/(n-1)) * SUM((yi-ym)2)).
are the standard deviations.
The x coordinate is for the Active measurements, and the y coordinate is
for the Passive measurements.
The Error = Avg(x) - Avg(y) / Avg(x);
You can click on the Node to view the time series and scatter plots, and
on the Type to get a space separated table of the data for that host.
We are currently
investigating in more
detail the differences in Passive vs. Active Measurements by using tcpdump.
Anomalies
Node (click for plots) | Active Tool (click for extracted data) | Passive | R | Samples | Avg(x) | Avg(y) | Std(x) | Std(y) | Error | min(x) | max(x) | min(y) | max(y) | Cov |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Node (click for plots) | Tool (click for extracted data) | Passive | R | Count | Avg(x) | Avg(y) | Std(x) | Std(y) | Error | min(x) | max(x) | min(y) | max(y) | Cov |