At 03:40 PM 12/19/01 -0500, Jeffrey Haas wrote:
>On Tue, Dec 18, 2001 at 05:49:07PM -0500, Kastenholz, Frank wrote:
>> Section 3.27 explictly says we preserve the current HbH model,
>> but others _may_ be added.
>
>A random thought:
>If we're considering non-hop-by-hop solutions, might not one also
>be able to elicit paths from the routing system where none existed
>before?
The requirements document would not prohibit such a scheme.
>How might this affect our requirements?
Personally, I don't see it as affecting them. My view is
that requirements set the minimum of what has to be
accomplished. If someone wants to propose a system
that does _more_, that's alright. Also, I see this
requirements document as outlining the general space
of the problem which needs solving and not as a hard,
cast-in-concrete list of things to do and not do. Someone
may come up with a "cool idea" that violates one of the
requirements. But that idea might make a whole bunch of other
things _sooooo_much_better_ that the tradeoff is worth it.
These tradeoffs would have to be evaluated and made by
whoever makes decisions on adopting the architecture
(presumably the IETF). But these are just personal
opinions.
Frank Kastenholz
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:03 EDT