Re: [sip-overload] Issue: Multiple algorithms -> Single algorithm for overload control

["Vijay K. Gurbani" <[e-mail]>]

On 06/16/2011 10:28 AM, [e-mail] wrote:
> Vijay,
>
> Thanks for the advice. I was indeed intending it to be an independent
> document, but a copy incorporating revisions to the original so that
> the differences would be made as clear.

Phil: Ah, OK.  Perfect, then.  More inline.

> Although in principle both approaches would support (or evolve to
> support) the same capability, I am concerned (if not convinced) that
> proceeding with a single mandated algorithm would result in
> non-adoption of the other algorithms by suppliers and in effect the
> alternatives would become redundant (rather like a piece of DNA with
> redundant sequences?).

Survival favours the fittest, so I strongly suspect that if one
algorithm does not perform as expected, it will be replaced by the
other.

That said, I think we have to be realistic here in recognizing that
the relative performance of all the algorithms is about the same.
So therefore, if the relative performance is the same then there
may be certain network deployments where the choice of one algorithm
seems better than the other.

Rate-based one will work well in managed networks where the set of
upstream clients sending requests is bounded and rather static; this
also makes it easier to apply capacity guarantees.  Loss-based one
works well where the set of upstream clients sending traffic is
unbounded and frequently changing.

I suspect that we will have a larger discussion on which one of
these becomes mandatory-to-implement, or whether it makes sense
implementing both of them (since no one is championing the windows-based 
one, I will keep it on the back-burner for now).

Thanks,

- vijay
-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / [e-mail]
Web:   http://ect.bell-labs.com/who/vkg/