North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: bring sense to the ietf - volunteer for nomcom

  • From: Jon Saperia
  • Date: Tue Sep 05 08:11:05 2000

on 09/05/2000 7:37 AM, Masataka Ohta at [email protected]
wrote:

> Sean;
> 
>> I think the IETF is valuable, but what do you tell investors when they
>> ask what's in it for them?
> 
> You have no problem, because it is as good as ISO.
> 
> Moverover, within NANOG context, it is better than ISO, because it
> is US-centric that 2 of 3 meetings in a year is held in US (remaining
> one often in CA).
> 
>> If UUNET needs some operational feature in a protocol, they
>> call up their Cisco engineer and say jump.  Presto, in the next release
>> train, feature X shows up. Who needs rough consensus?
> 
> Then, no one.
> 
> In theory, internet/routing areas are the only area where so valued
> rough consensus and interoperability could be meaningful.
> 
> Physical/datalink layer protocols are purely local. Transport/application
> layer protocols are chosen by the market, because of the end to end
> principle,

My comment for all these points:

    1. The cost of deployment, management and increasing complexity suggest
that common standards will reduce cost of operations for ISPs - why the
investors should be interested. They will say that the cost of the equipment
is only a fraction of the total cost of ownership with human labor being an
increasing component.

    2. It is true the ISPs and large customers have always had leverage with
various vendors - a good thing. A side effect is that this tends to divide
the market and hamper interoperability and management. Even when the market
is dominated by a few or just two vendors, there is division. If one is
dominant the others are at a disadvantage leading to one force that is hard
to steer even by large customers.

If, on the other hand, standards are well crafted with informed input by all
parties; vendors, users and other technologists, the result will be better
and operational costs will be reduced.

BTW, physical and datalink protocols are no longer purely local. Long haul
standards for various optical techniques are important unless one wants to
build a network out of only one vendors equipment.

/jon