Paying our way

[William Allen Simpson <[e-mail]>]

To the nay-sayers, part of the problem with IETF is how we pay our way.
I see this as a critical issue in our relations with ISoc, so this
discussion is pertinent.

Also, InterOp is part of our history.  One of the members of the IAB
sits on that body due to his involvement with InterOp.

> From: "Vinton G. Cerf" <[e-mail]>
> NSF, DOE, NASA and DARPA have provided support for IAB and IETF
> functions for many years, and for the IETF secretariat, specifically,
> since about 1988. In my earlier message, I mentioned that these
> agencies would like to see non-federal funding begin to take up
> the burden, and I believe this to be a very reasonable position.
>  
I absolutely agree.  Particularly as our efforts expand internationally.

So, how do *YOU* think that should happen?  What's your funding model?

> From: "Ole J. Jacobsen" <[e-mail]>
> Since I work for Interop, I'd really love to hear why you want us
> "replaced". I'd also love to hear where you get your numbers from,
> with that kind of accounting I could really do miracles, both in
> the US and elsewhere.
>
I meant operate an "educational" show for Internet vendors, probably run
by volunteers (like InterOp was), which generates funds to operate the
Internet backbones and staff.  My guess is that would make your show,
and several others, obsolete.

My InterOp literature says it's a non-profit, which I don't believe,
since a non-profit cannot be sold, and your books would be required to
be open for review by the public.

A private foundation/trust (which also cannot be sold) would not require
its books to be open, but would be required to disburse its assets (15
percent each year?), which I haven't heard as an attribute of InterOp
either.

My guess is InterOp generates about $20 Million gross per annum.

----

On the other hand, for the US backbone, NSF puts in about $11M per year,
plus $1M from Michigan (which ended this year).  MCI and IBM put in a lot
more (but are allowed to sell the "excess capacity" to commercial users).

The new NREN solicitation scheduled $6M, decreasing to 3 over a five
year period.  Note that this is *much* less than we get now!  The total
with 100Mb "high speed" may be $10M aggregate.

This reduction doesn't kick in until April 1994, but after that we're
going to be hurting for money.

According to Eric Aupperly at Merit, the dominant cost for running the
NSFnet is for physical infrastructure (actual lines).  Routers are
inexpensive relative to bandwidth.

The cost of staffing is also bandwidth insensitive, but rises as the
complexity of the net increases.  They've been able to keep staff
expansion at only a fraction of the increase in network connections
(which is currently exponential).

----

So, I was wrong.  An InterOp-like show wouldn't be able to fund the
actual backbone circuits world-wide.

But it would be able to fund the IETF and IRTF, including the editor(s),
administrative staff, the NIC, world-wide GIX centers, and possibly the
backbone routers.

I think that's a significant thing for us to do.  The question is, does
it fall within the ISoc Charter?

[e-mail]