Re: [conex] Crediting [was: Re: Review of draft-ietf-conex-abstract-mech-06]

[David Wagner <[e-mail]>]

Hi Bob, 

On Monday 15 July 2013 13:32:25 Bob Briscoe wrote:
> David,
> 
> I ought to read your credit draft first, but I'll just quickly 
> respond here (altho nothing much additional worth saying I'm afraid), ...
> 
> At 18:44 10/06/2013, David Wagner wrote:
> >Hi,
> >
> >I think there are some strong arguments for "aging" credits:
> >1) they avoid problems with restarted audits / rerouting: if credits 
> >are valid for e.g. 1 hour, any newly started auditer should not 
> >indicate reliable information for that time.
> >This can be extended to running auditers seeing new (rerouted?) IP 
> >addresses, but of course there is a trade-off for coverage.
> 
> This doesn't really "avoid problems". It just means the auditor knows 
> it has a problem for a certain period.
Well, that might be enough anyway. If the auditor knows it shouldn't penalize flows for a certain destination network for the time x, auditing could still cover most of the traffic and thus still be effective. 
> 
> >2) the value for credit is likely to degrade over time for the 
> >sender, typically already sent credits will count less than credits 
> >to be sent now.
> 
> That's not an argument for time-degrading credit, you're just saying 
> it's likely. It is likely if there's an argument for it, and it's not 
> likely if there isn't.
Agreed. But there are other & stronger arguments in favor of time-degrading / expiring credits. 
> 
> >This applies for rate-based credit / ConEx allowance mechanisms 
> >since, if using non-vanishing credits, the starting phase is much 
> >more costly than maintaining a flow. In other words: non-vanishing 
> >credit is an incentive to keep connections open. Do we want that?
> 
> We want the amount of credit to reflect the risk of cost to others. 
> (See my point in the email to Mirja about the perverse incentives 
> that would arise if you knew you would get your insurance premiums 
> back if you hadn't made a claim by the end of the year. Then it would 
> no longer be insurance, it would be an accident damage pre-payment 
> account.) But we also have the dilemma I outlined in the mail to Mirja.
> 
> 
> >Anyway, I don't yet have a good credit concept.
> 
> Yes, this is a problem.
I think this is a fundamental one since it questions the credibility of ConEx info and thus the incentive to deploy it. 
> 
> >Which also needs to address handling loss of ConEx-marked packets, 
> >at the sender and at the audit.
> 
> I don't think of that as a problem. I may not have covered it at the 
> IETF, but I think I did in my PhD thesis.
oops, I didn't check that. 
I wrote some sentences on it in the discussion draft, mainly coming to the conclusion that an auditor could estimate average loss of connection, thus providing an upper bound for loss of Conex-marked packets. 

Anyway, I'd really like to discuss ConEx crediting further. 

David 
> 
> 
> 
> Bob
> 
> 
> >David
> >
> > > Hi,
> > >
> > > so back on this one:
> > >
> > > > >9) Section 5.5.1 introdues the credit concept. Not sure if the 
> > meaning of
> > > > >credits is well enough specified here. My personal option is 
> > that credits
> > > > >should somehow get invalid (at some point in time). This is left open in
> > > > > the current text.
> > > > >
> > > > >
> > > > >I think we need to agree before we can talk
> > > > >about writing down what we agree...
> > > > >
> > > > >
> > > > >I think that abstract-mech needs to embrace
> > > > >*both*, explicitly if not implicitly.  I need to
> > > > >think about this some more, but I suspect that
> > > > >it means we have unnecessarily over constrained audit here.
> > > >
> > > > [BB]: We need to allow multiple experiments at
> > > > this experimental stage. But ultimately, sources
> > > > will need to unambiuously know what Credit means,
> > > > so they know how much to introduce and when.
> > >
> > > Yes, but we need to propose a mechanism when to send credits for 
> > the TCP mod
> > > draft and this means we need to have a common understanding how to handle
> > > credits in the endsystem and the audit. I guess that's what standards are
> > > good for. We might need a separate document for this. Not sure we 
> > are able to
> > > agree on this right now. As an alternative, I could also add some 
> > text in the
> > > TCP mod draft that the crediting is an open issue for experiments...?
> > >
> > > >
> > > > >Rather than thinking of Credit expiring after a
> > > > >time, one can think of the combination of recent
> > > > >Re-Echo signals and earlier Credit signals
> > > > >keeping the Credit state fresh within a flow. As
> > > > >long as you've sent Credit before a round of
> > > > >congestion, then if you send Re-Echo afterwards
> > > > >the Auditor can switch it round as if you sent
> > > > >the Re-Echo before and the Credit after.
> > >
> > > I don't think this would change anything. Maybe make it even worse. As you
> > > could also just send credit instead of ConEx marks and moreover there is
> > > still no incentive to send further marks when you have build up a large
> > > number of credits during Slow Start.
> > >
> > > > >
> > > > >So, when the Auditor holds Credit, it allows up
> > > > >to that amount of Re-Echo to be considered as
> > > > >having been sent before the congestion, rather
> > > > >than after. Then, as it switches the Re-Echoes
> > > > >back in time, it switches the Credits forward, so they always 
> > stay recent.
> > > > >
> > > > >Credit is primarily a mechanism to ensure the
> > > > >sender has to suffer some cost before it is
> > > > >trusted to pay back some cost. Credit doesn't
> > > > >need to degrade over time if the cost to the
> > > > >sender of introducing credit doesn't degrade over time.
> > > > >
> > > > >Does this move us forward, or do you still
> > > > >disagree? If so, I suggest a new thread would be useful.
> > >
> > > I have two concerns:
> > > 1) As mentioned above if a sender has sent a large number of 
> > credits during
> > > Slow Start and causes only few congestion during the rest of the 
> > transmission
> > > (as today's TCP usually does), there is no incentive to send further ConEx
> > > marks at all (neither credits nor loss/ECN ConEx marks).
> > > 2) When sufficient markings has been sent during Slow Start, no further
> > > credits should be needed. But if the audit for any reason will loose state
> > > (e.g. because of memory restriction or a new audit is used due to 
> > rerouting),
> > > the sender will still not send new credits as will and thus will cause the
> > > audit penalize the flow eventhough the sender did do nothing wrong.
> > >
> > > > >
> > > > >
> > > > >This is probably correct, but I really don't think it belongs in A-M.
> > >
> > > We might need an own document but there might also be some additional
> > > requirements that should be added to this document.
> > >
> > > >
> > > > [BB]: I don't think it should either. This is a
> > > > discussion with Mirja, rather than a proposal for text.
> > >
> > > _______________________________________________
> > > conex mailing list
> > > [e-mail]
> > > https://www.ietf.org/mailman/listinfo/conex
> > >
> >
> >
> >--
> >Dipl.-Inf. David Wagner
> >Institute of Communication Networks and Computer Engineering (IKR)
> >University of Stuttgart
> >Pfaffenwaldring 47, D-70569 Stuttgart, Germany
> >
> >web: www.ikr.uni-stuttgart.de   email: [e-mail]
> >phone: [phone]        fax: [phone]
> >-------------------------------------------------------------------
> 
> ________________________________________________________________
> Bob Briscoe,                                                  BT 
> 
>