[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: bug



> 
> 
> On Thu, 28 Jan 1999, Wenjie Weng wrote:
> 
> > Hi, we use ns-allinone-2.1b3 version. We found a bug in red.cc for
> > ECN simulations. When average queue size exceeds maxinum threshold,
> > the RED gateway drops packets. For ECN simulation, it should mark
> > the packet instead of dropping it.  
> 
> The ns code is right.
> RED should drop the packet; it marks only (instead of
> dropping) when between maximum and minimum threshold.
> 
> cheers,
> jamal
> 
> Computing Technology Labs (CTL), Nortel
> 
Hi, 

let me just try to clarify the matter a little.  First of all the
term "bug" is probably unlucky.

What we were trying to point out, is that in RFC2481, there is no hint
(correctly, that is concerned with the end hosts behavior, not with
router queue management) to how the router marks or drops packets,
although there is a clear indication that ECN should be used with RED
routers or other routers actively managing the queue. 

We run sume experiments and we could not obtain good results with ECN
unless the high threshold was set very high and we noticed that when
the queue was above the high threshold packets were dropped. 

We also noticed that, in an environmet with all ECN capable
connections, marking ALL packets above the high threshold leads to
good performance. 

What I would expect at this point, is that in a mixed environment a
RED router will take the decision on wheter to mark or drop the packet
based on its ECN capability and not only on the average queue length.

We just wanted to point out this fact and have more information (if
available) about RED queue management with ECN.

I hope I was reasonably clear. Let us know if we "missed" something in
the process. 

Renato Lo Cigno