Re: [ledbat] Fwd: New Version Notification for draft-ietf-ledbat-congestion-08.txt

[Mirja Kuehlewind <[e-mail]>]

Hi Arjuna,

heavy congestion does mean that for CTO time no data have been received. If 
data was received LEDBAT will compare the respective delay measurement with 
the base delay and increase or decrease the cwnd (except it is already 
MIN_CWND). Then if less packets are in flight than the cwnd allows it will 
send out a packet. Normally, if in CTO time no packet was received, it is 
assumed that all packets are loss and fightsize might be reset to 0. But an 
algorithm how to detect packet loss is independent from LEDBAT and thus not 
described in this draft.

I hope that helps...?

Thanks for your feedback!
Mirja


On Saturday 15 October 2011 10:37:06 Arjuna Sathiaseelan wrote:
> Dear Jana and others,
>
>  I went through the draft and am currently trying to figure out what
> the performance implications are with the latest draft through
> simulations (I have a few helpful UG students in India who have
> volunteered to do this for me - this includes modifying the current
> ns-2 code to reflect the latest draft). But unfortunately, I dont
> think we will be able to tell anything useful before WGLC ends!
>
> However I have one question ->
>  " A CTO is used to detect heavy congestion indicated by loss of all
>   outstanding data or acknowledgments, resulting in reduction of the
>   cwnd to 1 MSS and an exponential backoff of the CTO interval"
>
> So I am trying to see  whether heavy congestion actually means loss of
> "ALL" outstanding segments - so how does LEDBAT behave according to
> the current draft if atleast one packet went through? Just trying to
> understand.
>
> Regards
> Arjuna
>
> On 10 October 2011 00:51, Janardhan Iyengar <[e-mail]> wrote:
> > Dear all,
> >
> > A new version of the congestion control draft is in the repository. There
> > are two major mods in this revision:
> >
> > 1/ WGLC identified one major issue that needed to be addressed in the
> > LEDBAT congestion control draft -- LEDBAT response to extreme congestion
> > -- and we've tried to address that issue in this revision. We've added a
> > new mechanism, the Congestion Timeout (CTO), for a sender to respond to
> > extreme congestion. We do not specify how this should be implemented, but
> > we do note that a CTO can be implemented with or without a timer. In
> > terms of textual changes, we've added an update_CTO() function and a
> > branch for what to do if no acks are received within a CTO amount of time
> > in Section 3.4.2. We have changed the response to data loss to ensure
> > that a protocol, such as TCP, that uses the same timer for both
> > congestion control and for
> > retransmissions, changes its cwnd correctly.
> >
> > 2/ We have set the values for INIT_CWND to 4 and MIN_CWND to 2, and have
> > clarified the discussion of CURRENT_DELAYS and INIT_CWND/MIN_CWND in
> > Section 3.5.
> >
> > Please comment!
> > - jana
> >
> > --
> > Janardhan Iyengar
> > Assistant Professor, Computer Science
> > Franklin & Marshall College
> > http://www.fandm.edu/jiyengar
> >
> >
> > -------- Original Message --------
> > Subject: New Version Notification for draft-ietf-ledbat-congestion-08.txt
> > Date: Sun, 09 Oct 2011 16:33:33 -0700
> > From: [e-mail]
> > To: [e-mail]
> > CC: [e-mail], [e-mail],
> >  [e-mail], [e-mail]
> >
> > A new version of I-D, draft-ietf-ledbat-congestion-08.txt has been
> > successfully submitted by Janardhan Iyengar and posted to the IETF
> > repository.
> >
> > Filename:        draft-ietf-ledbat-congestion
> > Revision:        08
> > Title:           Low Extra Delay Background Transport (LEDBAT)
> > Creation date:   2011-10-09
> > WG ID:           ledbat
> > Number of pages: 19
> >
> > Abstract:
> >   LEDBAT is an experimental delay-based congestion control algorithm
> >   that attempts to utilize the available bandwidth on an end-to-end
> >   path while limiting the consequent increase in queueing delay on the
> >   path.  LEDBAT uses changes in one-way delay measurements to limit
> >   congestion that the flow itself induces in the network.  LEDBAT is
> >   designed for use by background bulk-transfer applications; it is
> >   designed to be no more aggressive than TCP congestion control and to
> >   yield in the presence of any competing flows when latency builds,
> >   thus limiting interference with the network performance of the
> >   competing flows.
> >
> >
> >
> >
> > The IETF Secretariat
> > _______________________________________________
> > ledbat mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/ledbat
>
> --
> http://about.me/arjuna.sathiaseelan
> _______________________________________________
> ledbat mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/ledbat



-- 
-------------------------------------------------------------------
Dipl.-Ing. Mirja Kühlewind
Institute of Communication Networks and Computer Engineering (IKR)
University of Stuttgart, Germany
Pfaffenwaldring 47, D-70569 Stuttgart

tel: [phone]
email: [e-mail]
web: www.ikr.uni-stuttgart.de
-------------------------------------------------------------------