Delayed ACK triggered by Header Prediction

[Noritoshi Demizu]

I am sorry for my delayed response.

I am sorry I was very ambiguous.  In my experience,
I transfered a small data (1 M bytes) without losses.

And the results are:

  1. from DragonFlyBSD to FreeBSD (HZ=1000): 1.2 Seconds
  2. from FreeBSD to DragonFlyBSD (HZ=1000): 2.0 Seconds
  3. from DragonFlyBSD to FreeBSD (HZ=10000): 1.1 Seconds
  4. from FreeBSD to DragonFlyBSD (HZ=10000): 1.5 Seconds

  (HZ represents the HZ value on the dummynet box on the path)

I made a web page of the result of my experiments at

  http://www.demizu.org/~noritosi/memo/2005/0318/

This page also shows the graphs of the packet traces captured
during experiments.

Thank you.

Regards,
Noritoshi Demizu