What's the differences between "quantum"s in htb and fq_codel qdisc?

by user762750   Last Updated July 12, 2019 14:00 PM - source

According to the manual page

HTB

quantum
how much bytes to serve from leaf at once {use r2q}

FQ_Codel

quantum
It is the number of bytes used as 'deficit' in the fair queuing algorithm. Default is set to 1514 bytes which corresponds to the Ethernet MTU plus the hardware header length of 14 bytes.

Questions:
1. Are they the same thing?
2. How to configure them to achieve best performance?



Answers 1


The essence of these parameters are same - how much bytes can be dequeued for single flow before switching to serving of other flow. For the best granularity this value should be near and little bigger than maximum packet size.

HTB has been implemented a long time ago and default r2q (rate to quantum) value isn't suitable for current speeds (you can see a lot of messages in the dmesg output about quantum of class ... is big. Consider r2q change). So better specify it manually with value of 1514 for unification.

Anton Danilov
Anton Danilov
July 12, 2019 13:31 PM

Related Questions




Load Dynamic Network Configuration Files

Updated February 28, 2019 09:00 AM


Linux adds the wrong default route

Updated September 14, 2015 16:00 PM