End-to-End QoS Network Design: Quality of Service for Rich-Media & Cloud Networks, 2nd Edition" Book.

from88 · ‎03-27-2016

Hello,

I'm learning about qos. Just read, that Scheduling (unlike queuing) occurs regardless of whether the interface is experiencing congestion.
But cant find more information about it. So it's a little bit strange for me, for example if i enable a legacy technology - custom queuing which uses round robin as a scheduler. Do when the congestion is absent this round robin approach will be used to decide how packets leave the interface ??

Thanks..

Peter Paluch · ‎03-27-2016

Hello,

I believe - and hope! - that Joseph Doherty, the resident expert on all things QoS, will join this thread but till he does, let me step in.

In my opinion, that statement is not correct. Scheduling, as far as I understand it, is the decision process that selects a particular queue from which the very next packet will be dequeued for further processing. At least for software queues, if there is no congestion on the physical interface, packets are not stored in software queues, which means that no scheduling takes place because there is nothing to schedule.

There are mechanisms that are active regardless of interface congestion - classification, marking, policing, shaping. However, queueing (as a process of putting packets aside into one or more [usually] software queues when the interface is congested) and scheduling (serving these [usually] software queues if they are nonempty) can, in my understanding, occur only when the physical interface is experiencing congestion.

Is it possible for you to quote a larger part of the text where that statement can be found? Is that a book, or an online resource?

Best regards,
Peter

from88 · ‎03-27-2016

Thanks fot answer, more or less i think the same. That quote is from "

End-to-End QoS Network Design: Quality of Service for Rich-Media & Cloud Networks, 2nd Edition" Book.

p.g 84

Congestion Management and Avoidance Terminology
Congestion management encompasses both queuing and scheduling. Conceptually, queuing
and scheduling are complementary but intertwined processes. These terms are quite
often incorrectly used interchangeably:
Queuing (which is also referred to as buffering) is the logic of ordering packets in
linked output buffers. Queuing processes are engaged when an interface is experiencing
congestion and are deactivated when congestion clears. As queues fill, packets
can be reordered so that higher priority packets exit the device sooner than lower
priority ones.
Scheduling is the process of deciding which packet to send next. Scheduling (unlike
queuing) occurs regardless of whether the interface is experiencing congestion.

Tried to search info on errata - don't find anything. I think Joseph Doherty is tired from my questions.. :)

Peter Paluch · ‎03-27-2016

Hi,

Yes, you're right, that statement is present in that book. I had a look but even seeing that statement in the full context did not help me to understand it better.

So this is what I did: I have e-mailed one of the authors asking for clarification, and I will make sure to post anything I will learn from his answer.

Best regards,
Peter

Joseph W. Doherty · ‎03-28-2016

Disclaimer

The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.

Liability Disclaimer

In no event shall Author be liable for any damages wha2tsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.

Posting

Scheduling (unlike queuing) occurs regardless of whether the interface is experiencing congestion.

This probably depends on how "congestion" is defined.

The book authors might consider "congestion" when an interface cannot keep up with packets' to-be-transmitted arrival rate, i.e. queues form and/or grow. Not congested might mean the interface is able to keep up with the arrival rate, i.e. queues don't form and/or queues shrink.

Queueing vs scheduling

End-to-End QoS Network Design: Quality of Service for Rich-Media & Cloud Networks, 2nd Edition" Book.