ietf
[Top] [All Lists]

RE: Thinking differently about names and addresses

2003-04-01 13:13:11
Egads. This list is still talking about the Identity Problem (i.e., that IP 
addresses are semantically overloaded in that they simultaneously indicate 
(network interface) routing topology and (node) identity). I just can't believe 
how we can continually talk about this problem and then not embrace solutions 
to it such as Bob Moskowitz's HIP (Host Identity Payload). 

All this reminds me of the famous Mark Twain quote where he said "Everybody 
continually talks about the weather but nobody does anything about it!!"

-----Original Message-----
From: J. Noel Chiappa [mailto:jnc(_at_)ginger(_dot_)lcs(_dot_)mit(_dot_)edu]
Sent: Tuesday, April 01, 2003 10:25 AM
To: ietf(_at_)ietf(_dot_)org
Cc: jnc(_at_)ginger(_dot_)lcs(_dot_)mit(_dot_)edu
Subject: RE: Thinking differently about names and addresses


    > From: "Tony Hain" <alh-ietf(_at_)tndh(_dot_)net>

    > your general perspective highlights the problem at hand. ..
    > the routing community believes the address is the topology locator,
    > while your & Dave's comments show the app community believes it is an
    > identifier.

To paraphrase Clint Eastwood (in 'Unforgiven'), "Belief's got nothing to do
with it".

The address field is absolutely required in order to get the packets to where
they are going. Therefore, it has to contain information that is organized in
such a way as to make the routing work.

If the characteristics of that information are such that it's not what the
application community want, they really don't have any other choice but to
find something else.

(I shouldn't need to point out that unless the routing works, the packets
won't get there at all, and if the packets won't get there at all, all other
debates are moot.)


    > I believe the multi6 discussion about creating a new identifier, to get
    > the app community to stop camping on the topology locator, will end up
    > creating a distributed database infrastructure almost identical to DNS.
    > We don't need two of those, so we should fix DNS.

This point was bounced around in the NSRG fairly extensively. I used to
believe as you did, that we could use the DNS as a namespace for naming
end-end entities. I got a lot of pushback on that from knowledgeable parties,
which I am sure they will be happy to rehearse (I'll leave it to them, as they
can do a better job of it); I have no religion about the answer.

I will also note that the DNS does "name" things other than end-end entities;
e.g. some individual names (for popular services) get translated into the
addresses of many different end-end entities.


    > I disagree with the perspective that subnetting or CIDR changed the
    > character of the address.

Absolutely - not at this level. It only had to do with how it was organized
to make the routing work (see above).

        Noel