[dtaht/deBloat] fa9d01: Still working through these.

[GitHub]

  Branch: refs/heads/master
  Home:   https://github.com/dtaht/deBloat
  Commit: fa9d011b6bc9531b9b235a438701aa4affe46252
      https://github.com/dtaht/deBloat/commit/fa9d011b6bc9531b9b235a438701aa4affe46252
  Author: Dave Taht <dave.taht at bufferbloat.net>
  Date:   2012-04-14 (Sat, 14 Apr 2012)

  Changed paths:
    A src/sfqnew.test
    A src/sfqnew_both.test
    A src/sfqold.test
    M src/simple_qos.sh

  Log Message:
  -----------
  Still working through these.

With the mark 1 eyeball, sfq with enqueue to head wins. It recovers
faster than the alternative when burdened by multiple streams.


  Commit: faac0d286fab0f0416272bc8171310c45bfd7bad
      https://github.com/dtaht/deBloat/commit/faac0d286fab0f0416272bc8171310c45bfd7bad
  Author: Dave Taht <dave.taht at bufferbloat.net>
  Date:   2012-04-14 (Sat, 14 Apr 2012)

  Changed paths:
    A src/sfqnewboth.log

  Log Message:
  -----------
  I show perfect fairness with 50 streams and 150 buffers...

this is with sfqred on both sides, htb on the far side doing
4Mbit limit (and we hit that, exactly), and no starvation.


Compare: https://github.com/dtaht/deBloat/compare/a79ec97...faac0d2