Re: [Dime] Review of

On Jul 25, 2013, at 9:59 AM, Janet P Gunn <jgunn6@csc.com> wrote:

> You might also want to take a look at
> 
> draft-ietf-soc-overload-control-13 
> 
> 
>  and
> 
> draft-ietf-soc-overload-rate-control-04
> 
> 
> While SIP is different from DIME, the SIP overload work was, in some senses, the "inspiration" for the DIME overload work, and some of the same issues were addressed. 
> 
> In particular, there was the same concern about "loss based" vs "rate based" throttling, resulting in two separate drafts.  Both use the same (in band)  communications protocol, and the same parameters, but use different parameter values , and different default algorithms. 
> 
> I agree that it is, in general, better  for the server to tell the client "do this", rather than "this is my status, you figure out what to do".  But DIME may not match the "in general" situation. 

It may also be worth pointing out that, at least in draft-roach-dime-overload-ctrl, you really don't have a case of "here's my status, figure it out" in general. Peers negotiate an algorithm, and that algorithm specifies what the client should do. The algorithms are extensible, so it would be possible to make one that fits the "sender figures it out" model, but there's no assumption of that.

For the loss-based algorithm included in the draft, the overload metric specifies a percentage reduction. For example, a metric of 50 means the client should reduce traffic by 50. While that still gives the client some leeway in determining which requests to drop, that's considerably different from saying "I'm 50% overloaded; figure it out." In fact, while we're sometimes imprecise here in conversation,  there's no assumption in that algorithm that "reduce your offered load by 50%" maps in any way to "I'm overloaded by 50%".

> 
> Janet 
> 
> 
> 
> dime-bounces@ietf.org wrote on 07/25/2013 08:50:06 AM:
> 
> > From: Alan DeKok <aland@deployingradius.com> 
> > To: Eric McMurry <emcmurry@computer.org> 
> > Cc: "dime@ietf.org" <dime@ietf.org> 
> > Date: 07/25/2013 08:50 AM 
> > Subject: Re: [Dime] Review of 
> > Sent by: dime-bounces@ietf.org 
> > 
> > 
> > > yes, Diameter clients are different.  They tend to be 
> > sophisticated servers doing a number of functions with Diameter 
> > being used for some of their control signaling.  Their other 
> > functions usually require the use of other protocols and they may 
> > have to consider information in making overload control decisions 
> > that servers do not have.  The requirements draft may be helpful for
> > describing the problem domain for this (http://tools.ietf.org/html/
> > draft-ietf-dime-overload-reqs-09)
> > 
> >   I scanned it.  I'll take a deeper look.
> >_______________________________________________
> DiME mailing list
> DiME@ietf.org
> https://www.ietf.org/mailman/listinfo/dime