Re: SIP Addressing Limitations

[Vince Fuller <[e-mail]>]

    > It amuses me greatly that there's anyone who still seriously
    > believes that network topology, and geography, are even remotely
    > related when it comes to international connections.
    >
    Apparently I am in good company.  With Rehkter, Li, and the CIDR
    luminaries.

I guess the fact that my name appears first on the CIDR document makes me a
"CIDR luminary", so I'd like to emphatically refute your assertion. CIDR is
basically a hack which attempts to provide, with minimal impact on Internet
hosts, better scaling of routing in a system which has horribly confused the
two distinct concepts of Name (that which uniquely identifies a communicating
object) and Address (that which describes where an object is located). The
techniques described in CIDR should most definitely NOT be used as an example
of a well-designed naming, addressing, and routing architecture.

If the best SIP/IPAE can come up with for dealing with the addressing/naming
problem is to be "just like CIDR", than the Internet is in real trouble in the
long run. If you read CIDR carefully, you will see that for it to work really
well, fairly constraints must be placed on the topology of the Internet (i.e.
most sites are singly-homed and provider mobility is relatively rare). As the
Internet topology becomes more complex and as entropy (i.e. network mobility)
is introduced into the system, the benefits of CIDR-style aggregation are much
diluted over time.

Until SIP/IPAE can produce a real solution to the "re-addressing problem"
(which applies to both geographic- and provider-based "addressing"), you can
spend multi-100's of hours on an "addressing" plan which simply won't produce
good aggregation in a poorly-constrained Internet topology.

	--Vince