0

In the HTB documentation can be read:

Note: HTB as child of another HTB is NOT the same as class under another class within the same HTB. It is because when class in HTB can send it will send as soon as hardware equipment can. So that delay of underlimit class is limited only by equipment and not by ancestors. In HTB under HTB case the outer HTB simulates new hardware equipment with all consequences (larger delay)

EDIT: My question is: the delay for having a qdisc children of another is large enough for not considering this approach over having only one qdisc?

The thing is that I must have enabled QoS because I want prioritize some traffic over other one, I'm just asking for the approach to do that.

Thanks in advance.

Diosney
  • 305
  • 5
  • 12
  • What's the question? Those two scenarios are completely different use cases. If you wouldn't want to have a larger delay for some of your packets, you hardly would turn to QoS, would you? – the-wabbit Oct 12 '11 at 16:36
  • __@syneticon-dj:__ Hi! Thanks for your time. I updated the post on this. – Diosney Oct 12 '11 at 19:03

1 Answers1

0

There are situations where you want delayed transmission - e.g. when simulating a slow link. In a simple case of prioritizing traffic, creating two or more HTB classes with different priorities and using filters to assign traffic to them would do. But nevertheless you should add the stochastic fair queuing qdisc as a child qdisc of your HTB classes to make sure that a single connection within a class does not starve the others. There are some examples at lartc.org on this topic - check them out.

the-wabbit
  • 40,319
  • 13
  • 105
  • 169