[e2e] New approach to diffserv...

[Melinda Shore]

At 12:27 PM 6/16/02 -0500, John Kristoff wrote:
>I can assure you that at least one institution I know of absolutely
>does want e2e transparency.  I've heard that others care about it
>too, but we at least felt strongly enough to write up our own rfc
>about security and its role in 'the network'.

I hope you understand that the economics driving networking
at educational institutions is not at all like the economics
driving networking at commercial service providers or the
economics driving networking for internal use at large
enterprises.  The assumption that increasing access adds value
to the network may be true for some network owners but 
certainly is not for other network owners.  For example, at
a company like Cisco, or Nokia, or Qualcomm, or ..., the
network provides value when it allows employees to exchange
information easily.  It adds value when it allows customers
and potential customers to have access to sales and support
material.  It *subtracts* value - a lot of value - when it
allows customers or competitors or any number of other categories
of unwelcome interlopers to have access to source code or
development documents.  Also, the funding model is different
and that matters a very great deal.

Or consider addressing.  We know from experience with the phone
companies that if they can treat addresses like property they will,
and they'll engage in unattractive behaviors to protect the value
of that property.  It should not have been a surprise when the 
same thing started happening with IP addresses, and it should be
less of a surprise that the consequences of that are pretty severe
when taken in the context of a protocol where addresses care a lot
of semantic weight.

None of this should be taken as an apology for middleboxes.  I
actually think that even among people who routinely say awful things
about NAT that it's not fully appreciated what a serious problem NAT
is for even modestly complex applications - there are *so* many
ways that NAT can interfere with an application.  At the same time
it's also very clear that people put those things in their network
to solve problems, and they aren't going to pull them out unless they're
given some other technically credible, manageable, and affordable way
to solve those same problems.  Declaiming the glories of end-to-end
transparency and telling the people who own the networks that they can 
run more services better if they'd simply uninstall their firewall just 
isn't going to do it.  We need more work like the Ioannidis paper on 
distributed firewalls (without Keynote, but that's a different matter) 
and Bob Moskowitz's HIP papers.  When all is said and done there's just 
too much saying and not enough doing.

Melinda