Measurement provides evidence. Analysis of evidence builds intuition especially when flawed or misleading evidence is uncovered.
Here we discuss a quantity that surges in traffic lulls. The context is recreational observation of digital ham radio traffic. We learn from our play.
From: Ward Cunningham To: Pat Cunningham Subject: Misleading Datalog Results
This nightly "blue bulge of inactivity" is an artifact of how I empty my buffer at a load-dependent rate. enlarge
I will add logic that empties 15 seconds of buffer late in the NTP synchronized cycle, a complexity I had not anticipated.
Datalog sampling has a steady drift of a few milliseconds each period (one minute in this case). This will lead to a beat signal from the unsynchronized clocks but one that is more theoretic than noticeable.
Things to notice.
- correspondence between participants - subject is diagnosis - informal name for phenomenon, blue bulge - identified cause, load-dependent interval - corrective action - further concerns