[Iccrg] Heresy recapped

[Lachlan Andrew]

Greetings Wes,

On 17/04/2008, Eddy, Wesley M. (GRC-RCN0)[VZ] <wesley.m.eddy at nasa.gov> wrote:
>
>  The from:
>  http://ee.lbl.gov/papers/congavoid.pdf
>  "While algorithms at the transport endpoints can insure the network
>  capacity isn't exceeded, they cannot insure fair sharing of that
>  capacity. Only in gateways, at the convergence of flows, is there enough
>  information to control sharing and fair allocation."
>
>  Both references describe the selection of AIMD without reference to
>  fairness, but specifically for stability.

I agree (now) that that wasn't the original intention.  However this
reminds me of US discussions of gun laws based on what the writers of
the constitution intended hundreds of years ago, when circumstances
(firepower, data rates) were vastly different.

Although I don't think it matters whether we get "exact fairness", I
do think that
1) congestion control has assumed the task of resource allocation on
*some* links, and
2) whatever allocates resources has a responsibility to ensure no user
is starved.

I agree users will not "compare notes", but a user who is being
starved (e.g., by excessively slow starting) will often be
dissatisfied, and as researchers we should try to find ways to avoid
that.  <shamelessPlug>For example, in
<http://netlab.caltech.edu/lachlan/abstract/dq_infocom99.pdf> we
looked at a scheme aiming to keep users satisfied at the expense of
short-term fairness.</shamelessPlug>

The current notions of fairness may be broken and/or overemphasised,
but I'm not convinced that it is a non-issue.

Cheers,
Lachlan

-- 
Lachlan Andrew  Dept of Computer Science, Caltech
1200 E California Blvd, Mail Code 256-80, Pasadena CA 91125, USA
Ph: +1 (626) 395-8820    Fax: +1 (626) 568-3603
http://netlab.caltech.edu/lachlan