> But if we want one client to be able to modify part of a RIB entry, or if
we
> want to be able to work with things other than RIB entries, then all of
the
> issues that were discussed several years ago come back into play.

First -- in terms of granularity -- I don't think anyone should be modifying
parts of a RIB entry. Rather, each "object" should be treated as an "object"
in full. This is something that needs to be addressed in the models -- what
is essential, and what is accidental.

Second -- I'm not arguing one problem should be solved, and others
shouldn't. I'm saying I think we might have two different interrelated
problem sets, both of which need to be solved. For instance, it might be
useful to say something like: "In the RIB, overwrites are not an error; in
all other areas they are." And then define the models and processes to
handle these two different cases. I don't know that using one unified way of
dealing with things is necessarily going to solve all the problems at hand. 

Recognizing this might help us move forward in a coherent way, rather than
stumbling over the same set of problems in different guises each time some
new point is undertaken.

:-)

Russ 

_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs

Reply via email to