RE: WG next steps

["Aldrin Isaac" <aisaac@bloomberg.com>]

% From: owner-multi6@ops.ietf.org [mailto:owner-multi6@ops.ietf.org]On
% Behalf Of Craig A. Huegen
%
% There is absolutely an interest in the network/routing
% based solutions.  It's just that the operator folks who
% will be using them significantly as they're deployed
% probably aren't in this forum (or are trying to keep
% up with billions of other things at the moment).
%
% It is undesirable for a large enterprise to use a
% multi-address/host based solution without a mechanism
% to interact with the network topology to perform path
% selection.  Network operators know more about the
% viability of paths (whether by technology or manual
% policy) than the hosts do.  Left-side longest-match
% is hardly a sufficient routing  protocol and/or policy
% control for a multi-provider topology like the Internet.
% The network *MUST* be a part of the solution --
% enterprises want policy control, and policy control
% should be centralized.
%

I strongly agree.

The network is a policy enforcement tools used by the enterprise
network manager to control cost and liabilities related to e-business.
This involves firewalls, ALGs, AND controls implemented using routing.
The enterprise network will continue to play this role whether anyone
likes it or not.  If you don't believe me, have a talk with the
network manager of any firm that takes any and all risk seriously
(even those *we* may qualify as non-risks).

BGP and TE protocols are critical routing control tools for the
implemention of an ISPs business agreements and policies.  The
enterprise needs the same, if not more, control at the routing level.
Especially to mitigate the burden of having to support multi-homed
hosts!  Ignoring this will result in certain "cool" features of IPv6
to become a major thorn in their side and may result in unwanted
"fixes" (ex: NAT).  The assumption that fine-grain route control by
the enterprise is unneccessary is very bad judgement.

I know of multiple cases for which IPv6 is a non-starter simply
because of the reduced role played by the [enterprise] network in the
source address selection process and the lack of source-based
site-exit routing.  These are IMHO critical basic componants to any
network-based or host-based IPv6 multihoming solution for the
enterprise.

The enterprise people are caught between the ISP oriented people on
one end exerting their influence on address delegation and the
application/host oriented people on the other end exerting their
influence on address selection.  One or the other (or both) has to
give a little to make some room for the guy in the middle.

-- aldrin