Alexandru Petrescu<petrescu@crm.mot.com> wrote:
>Alex Zinin <azinin@nexsi.com> writes:
>> 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...
>
> Following this line, is it possible to imagine an
> architecture where addresses are topologically significant
> and prefixes are also topologically significant? And then
> having a mechanism for name-to-prefix mapping?
>
> I might be far off from what "topologically significant"
> means exactly.
Well, there are is a wide range of solutions proposed for
mobility. Some of them involves some network layer tunneling
or per host routing stuff to hide mobility from correspondent
hosts. And most of them involves some form of end-to-end
signalling to notify correspondents about mobility (at vaious
layers, e.g. IP layer in Mobile IP, transport layer in STCP
or application layer in SIP mobility).
Of course, both of the above approaches have their limits.
End-to-end signalling in a global scale may introduce
unwanted amounts of location updates and result in high
delay. If mobility is handled by the routers then either sub-
optimal routing takes place (routing via home agents) or
mobility is limited (e.g. to micro-mobility domains) due to
scalability limits.
An alternative might be a combination, where end-to-end
signalling is used, but the routing infrastructure (or layer
2) hides some local movements to limit end-to-end signalling.
In such an architecture, addresses are locators, but those
locators do not change frequently and may perfectly serve to
identify a host during short lived transactions. The more
movements the routing infrastructure hides, the less end-to-
end signalling is needed. For stationary hosts, locators can
practicaly be used as identifiers for most transactions
(except for the rare case of renumbering).
It is an open question on which level do we need end-to-end
mobility signalling.
Zoltan
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:04 EDT