Nonlinear RED implementation in ns-3

Nonlinear RED (NLRED) [1] is a variant of RED in which the linear packet dropping function of RED is replaced by a nonlinear quadratic function.

This patch provides implementation of Nonlinear RED in ns-3 along with documentation, test-suite and example program.

Your reviews are much appreciated.

Thanks,
Nichit Bodhak Goel

[1] Zhou, K., Yeung, K. L., & Li, V. O. (2006). Nonlinear RED: A simple yet efficient active queue management scheme. Computer Networks, 50(18), 3784-3794.

[Stefano Avallone, 2017-01-27 14:10:57]

Inline are a couple of questions related to how p is computed.

Also, I would like to ask you to move the example to examples/traffic-control
(due to dependency on point-to-point, internet and applications).

Thanks.

https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
File src/traffic-control/model/red-queue-disc.cc (right):

https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
src/traffic-control/model/red-queue-disc.cc:712: 
I can't find in the reference that maxP is set to 1.5 times its old value. Could
you please point me to the place where it is stated?

https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
src/traffic-control/model/red-queue-disc.cc:726: }
I think this can be further simplified:

  else
    {
      /*
       * p ranges from 0 to max_p as the average queue size ranges from
       * th_min to th_max
       */
      p = vA * qAvg + vB;
      if (m_isNonlinear)
        {
          p *= p; (or p *= p * 1.5);
        }
      p *= maxP;
    }

[PhaniKiran Somanchi, 2017-01-29 19:45:08]

On 2017/01/27 14:10:57, Stefano Avallone wrote:
> Inline are a couple of questions related to how p is computed.
> 
> Also, I would like to ask you to move the example to examples/traffic-control
> (due to dependency on point-to-point, internet and applications).
> 
> Thanks.
> 
>
https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
> File src/traffic-control/model/red-queue-disc.cc (right):
> 
>
https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
> src/traffic-control/model/red-queue-disc.cc:712: 
> I can't find in the reference that maxP is set to 1.5 times its old value.
Could
> you please point me to the place where it is stated?
> 
>
https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
> src/traffic-control/model/red-queue-disc.cc:726: }
> I think this can be further simplified:
> 
>   else
>     {
>       /*
>        * p ranges from 0 to max_p as the average queue size ranges from
>        * th_min to th_max
>        */
>       p = vA * qAvg + vB;
>       if (m_isNonlinear)
>         {
>           p *= p; (or p *= p * 1.5);
>         }
>       p *= maxP;
>     }

At the end of section 3.1 in the reference it is stated why maxP is 1.5 times
old maxP
Thanks.

[Nichit Bodhak Goel, 2017-01-31 11:05:35]

Hi Stefano,

Thanks a lot for reviewing the code and providing suggestions. I have uploaded a
new patch with the suggested changes.

Please let me know whether the new patch looks fine.

Thanks,
Nichit Bodhak Goel

https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
File src/traffic-control/model/red-queue-disc.cc (right):

https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
src/traffic-control/model/red-queue-disc.cc:712: 
On 2017/01/27 14:10:56, Stefano Avallone wrote:
> I can't find in the reference that maxP is set to 1.5 times its old value.
Could
> you please point me to the place where it is stated?

In the paragraph just before Section 3.2, authors have explained the reason of
using 1.5 * maxP. Please let me know if any more information is required.

https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
src/traffic-control/model/red-queue-disc.cc:726: }
On 2017/01/27 14:10:56, Stefano Avallone wrote:
> I think this can be further simplified:
> 
>   else
>     {
>       /*
>        * p ranges from 0 to max_p as the average queue size ranges from
>        * th_min to th_max
>        */
>       p = vA * qAvg + vB;
>       if (m_isNonlinear)
>         {
>           p *= p; (or p *= p * 1.5);
>         }
>       p *= maxP;
>     }

Done.

However, I have slightly modified the code which you suggested so that a new
reader can relate it easily with the paper.

Please let me know if these changes are fine.

[Nichit Bodhak Goel, 2017-03-16 09:00:52]

On 2017/01/31 11:05:35, Nichit Bodhak Goel wrote:
> Hi Stefano,
> 
> Thanks a lot for reviewing the code and providing suggestions. I have uploaded
a
> new patch with the suggested changes.
> 
> Please let me know whether the new patch looks fine.
> 
> Thanks,
> Nichit Bodhak Goel
> 
>
https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
> File src/traffic-control/model/red-queue-disc.cc (right):
> 
>
https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
> src/traffic-control/model/red-queue-disc.cc:712: 
> On 2017/01/27 14:10:56, Stefano Avallone wrote:
> > I can't find in the reference that maxP is set to 1.5 times its old value.
> Could
> > you please point me to the place where it is stated?
> 
> In the paragraph just before Section 3.2, authors have explained the reason of
> using 1.5 * maxP. Please let me know if any more information is required.
> 
>
https://codereview.appspot.com/315400043/diff/1/src/traffic-control/model/red...
> src/traffic-control/model/red-queue-disc.cc:726: }
> On 2017/01/27 14:10:56, Stefano Avallone wrote:
> > I think this can be further simplified:
> > 
> >   else
> >     {
> >       /*
> >        * p ranges from 0 to max_p as the average queue size ranges from
> >        * th_min to th_max
> >        */
> >       p = vA * qAvg + vB;
> >       if (m_isNonlinear)
> >         {
> >           p *= p; (or p *= p * 1.5);
> >         }
> >       p *= maxP;
> >     }
> 
> Done.
> 
> However, I have slightly modified the code which you suggested so that a new
> reader can relate it easily with the paper.
> 
> Please let me know if these changes are fine.

Hi Stefano,

I have uploaded the new patch (Patch 3) according to the updated version of
ns-3-dev.

Thanks,
Nichit Bodhak Goel

[Stefano Avallone, 2017-05-28 11:12:38]

Pushed to ns-3-dev with changeset 12891:3b239b1c038c.

Thanks|

PS Could you please close this code review?