The whole thing (of not even allowing
static addressing -- only topologically
significant, totally dynamic one) was
explicitly proposed first time (up
to my knowledge) in the Antonov's TRAP:
http://gato.kotovnik.com/~avg/old_page/draft-trap.txt
(there are even protocol spec "IDs"
in the same dir). But then, he's also
strange, isn't he? :)
But then again the aggregation question
comes up. The natural thing to do next
would be to dynamically aggregate (which
was also proposed in TRAP assuming
the hierarchical Internet topology),
and this would be fine (subject to
policy, etc.) if the assumption
was true -- the question about how
you'd aggregate beyond the AS level
given the current statistics and trends
remains.
-- dima.> -----Original Message----- > From: Tony Li [mailto:tli@procket.com] > Sent: Wednesday, April 10, 2002 1:49 AM > To: Alex Zinin > Cc: irtf-rr@puck.nether.net; avri > Subject: Re: mobility > > > > > In my mind, that would be the preferred solution. > > The address should always be topologically significant and the host > identification should be orthogonal to the topological info. > > But then, I'm strange... > > Tony > > > Alex Zinin writes: > | > | Following up on this old thread... > | > | I'm a little worried about the "MUST support" wording > | here. Specifically, one could envision an architecture, > | where mobility is absolutely transparent to the routing > | system by making address be always topologically signi- > | ficant and having a separate mechanism for name-to-address > | mapping... > | > | Am I worrying too much here? > | > | Alex > | > | Tuesday, March 5, 2002, 8:24:59 AM, avri wrote: > | > | > in ngarch-req: > | > | > > There are two kinds of mobility; host mobility and > | > > network mobility. Host mobility is when an individual > | > > host moves from where it was to where it is. Network > | > > mobility is when an entire network (or subnetwork) > | > > moves. > | > | > > The architecture MUST support network level > | > > mobility. > | > | > does the absence of statement about requiring host mobility > | > indicate that this is not something the architecture should > | > be required to support? > | > | > i think the architecture should support both. > | > | > i am also curious, you state that the two are definitely two > | > different kinds of things. it seems to me that there may > | > be architectural abstractions in which 'user' mobility is a > | > special case of 'network' mobility. i think any future > | > architecture needs to support both, and if it can do so > | > through a common method, all the better. > | > | > a. > |
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:04 EDT