Re: [sip-overload] draft-ietf-soc-overload-control-02 and algorithm agility

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

On 03/29/2011 05:50 PM, Hadriel Kaplan wrote:
> But that to me argues we shouldn't pick the loss-based one to begin with.
> Either they're all effectively the same, or they're not. As far as I
> know, the results of the design-team were they were the same, so the
> loss based one was picked just because it's the simplest to implement.
> If a rate-based is instead more effective, even if it's just for "corner
> cases", then we should pick that one instead.

The rate-based one requires more state in the server.  Keeping
less state in the server is an argument to pick loss-based as the
default algorithm instead of rate-based.  The latter requires the
server to assign --- and keep track --- of a share of its overall
capacity with each upstream client.  There is a good discussion
of this in the design document [1], Sections 9.1 and 9.2.

[1] http://tools.ietf.org/html/draft-ietf-soc-overload-design-05

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/