At 02:08 PM 12/11/01 -0800, demir wrote:
>Isn't the concept of intra-domain/inter-domain split is "hierarchical" and
>addressed in:
>Paul F. Tsuchiya, "The Landmark Hierarchy: A New Hierarchy for Routing in
>Very Large Networks", Proceedings of ACM Sigcomm 1988, pp. 128-134.
What we'd like to do is eliminate the notion that there
is this thing called inter-domain and it is fundamentally
different from intra-domain, that there _must_ be 2
and exactly 2 such levels, that they _must_ use different
protocols, and so on.
Conceptually, we'd like to see the routing model allow for
a hierarchy of elements. What is internal to one element
(sort of like intra-domain) is in fact external (more or
less like inter-domain) to the elements contained within
that element. Elements at different levels of this hierarchy
might make different 'information' known externally vs internally,
might pass around different policy, etc, etc. Even if these
levels are all within the same administrative control (aka
domain/AS/...) I realize I'm talking a bit about a possible
solution to meet the requirements rather than the requirements
themselves, but I'm doing it as an example, trying to make
the requirement a bit more clear (I hope!)
On Tue, 11 Dec 2001, Sean Doran wrote:
>> this split is "institutionalized"? I am sure the authors and
>> editors will be happy to think of definitive ways to exclude
>> architectures which ELIMINATE the concept of intra-domain/inter-domain
>> protocol splits.
The editor heartily concurs (assuming that the rest of the RRG
concurs with the idea!)
frank kastenholz
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:03 EDT