ietf
[Top] [All Lists]

Re: WG Review: Open Pluggable Edge Services (opes)

2001-06-19 16:40:03
I think it does matter - there's an idiom to IP networking,
and if what you do on top of IP isn't idiomatic things are
inevitably going to be clunky and, at worst, break
completely.  That said, I don't think it's accurate that
the cause of all of this unidiomatic crap in the network
is the result of lack of understanding (or outright
stupidity) on the part of those doing the crap.  

not all of it, but some of it does demonstrate a profound
lack of foresight.

I think
the problem is that with the commercial success of IP the
economic model has changed and the incentives have changed.

and the models continue to change, and there's no end in sight
yet.  I think we are in a period of experimentation with 
many different kinds of models to see which ones work best.
It's not at all clear where we will end up, because the models
make assumptions about an economy that is itself changing.

to take a crude exmaple, if banner advertisements continue
to drop in value, business models based on insertion of
banner ads that seem viable today may not seem viable in 
six months' time.

(and there's only so much space in a browser window , and
only so much of a user's attention to be borrowed. someone
is going to have to figure out to whom it belongs :)

We've always behaved as if the notion that the value of
the network is a function of the number of connected nodes
is true.  While it may very well be, I think we now need
to start raising the question of value to whom and
whether or not there might be other considerations.  It
may be the case that my network is more valuable to me if
I have access to you, but that it's less valuable to me if
you (and you and you and you) have access to it.

that's a good way of expressing it.

Because enterprises want to protect their property and
service providers want to be able to make money with their
services they're asking for the means to make the edges of
their networks visible.  They're working from a business
model that requires it.  Many of the middleboxes we're
arguing about are instantiations of policy, and policies
are instantiations of business models.  It seems clear to
me that if we want to restore end-to-end networking we need
to find alternate business models.  Finger-shaking won't
solve the problem.

I don't think that IETF is compelled to support arbitrary business 
models - especially those that don't make good engineering sense.  

It seems that business models are a bit like religions. You should 
be able to believe whatever you choose, but that doesn't mean you
can realistically expect the rest of the world (or anyone for that 
matter) to conform to it.  And a religion is not a justification to 
harm things that work for others

Keith