Re: [RAM] First cut at routing & addressing problem statement

Hi,

I have not had time to review the full document yet, but I do have some 
immediate comments.

In Section 3, 4th para, we *do* build routers to meet today's 
requirements and so it is unnecessary to ask a question that has already 
been clearly answered in the affirmative.

In Section 3.1, first para, the wording here leaves out the most 
important "must", which is that routers must process changes to the 
network topology.  Moreover, it leaves as quite murky as to whether the 
problem is in the forwarding path, the update path, or both, or whether 
this is dependent on design architecture.  I would personally like to 
see some performance numbers here.

In Section 4.3 first para, last sentence, given current discussions on 
RAM, I would make the following change for clarity:

OLD:
> However, each individual PI
>    prefix must be propagated throughout the DFZ and adds to the DFZ
>    routing load.

NEW:
> However, with the current BGP-based routing system, each individual PI
>    prefix must be propagated throughout the DFZ and adds to the DFZ
>    routing load.

Section 4.4 should reference (informationally) RFC 1627, where we first 
raised this concern 13 years ago.

Finally, in Section 6, I think we should add that it is desirable that 
those who multihome or make use of traffic engineering incur whatever 
associated costs.  This follows in part from your earlier business 
alignment discussion.

Eliot

_______________________________________________
RAM mailing list
RAM@iab.org
https://www1.ietf.org/mailman/listinfo/ram