Re: [rrg] Constraints due to the need for widespread voluntary adoption

[Dae Young KIM <[e-mail]>]

On Fri, Dec 4, 2009 at 8:23 AM, Noel Chiappa <[e-mail]>wrote:

>
> Since "address" to most people does mean a 'name' (in the generic sense of
> "name") with some location information in it, I am going to assume that
> your
> "ID" means a name which is location dependent.
>

No, my 'ID' means a name which is location 'independent'.

>
> That means the "ID" will have to change if the host is attached to the
> network in a different place; and the experience of the past decade or two
> is
> that this is not acceptable to the users. They want something _in addition_
> to DNS names which does not have to change when the host is attached in a
> new
> location.
>

No, in one preferred model, ID does not change along the networks. If I'd
try to stick to the 'usual usage of terms' in this community, only the
'locator' changes according to the network which they attach to.

>
>    > ID -> locator; this is not the job of Internet layer. In this context,
>    > the locator here is to me the address of the underlying layer,
>
> To most of us, 'locator' means a 'name' with global semantics; i.e. it can
> name a location anywhere in the internetwork, and it can be understood
> anywhere in the network, and things everwhere will interpret a given
> locator
> to mean the same location.
>

Here, we're clearly apart. To me, locator need not be globally unique. It
can change across networks. Each 'enterprise network' may use its own local
locator name space. At each network boundaries through routing, you just
change to a new locator in the new (transit or stub) network. All along,
however, your ID will remain constant.

>
> Redefining terms to your own private definition is not very useful for
> clear
> communication.


OK, agree. Will refrain from doing that as long as I'm discussing in this
community.


> Terms that mean what you seem to mean here are things like
> 'MAC address' or 'physical network address'.
>

Yes. In fact, now that you have an ID, there's no need to have something
like 'PoA' address separately in the same layer. You just rely on your
underlying layer to deliver your packet through it's underlying layer to my
next hop in my routing table. So, all I need is

     - mapping from my (source) ID to the subnetwork address (or MAC addr or
physical addr according to your examples)

     - and also mapping from the ID of my next hop in my 'internet layer'
routing table to the subnetwork address (MAC or physical address) of that
very next hop

     - at reception of my packet, my next hop (router) will look into the ID
of my intended final destination ID and decide to which next hop (node or
router) the packet has to be relayed

      - and the whole process above repeats.

>
>    > To me, they're one and the same thing. There's no routing without
>    > identifying your partner. There's no identifying your partner without
>    > routing to, i.e., without locating it. The two are one and the same.
>
> Ah, no. I am "J. Noel Chiappa" no matter where I am, and you can identify
> me without knowing either i) where I am, or ii) how to get there.
>

OK, you're talking about DNS name(or URI?). Noel is the 'DNS' name which, of
course, is unique independently of whatsovever.

Then, we said DNS name should be mapped to ID. (Noel to 234 Madison Ave.).
There's no routing (or better said locating?) without identifying your
house. There's no identifying your house without successfully locating it.

>You might be interested in reading IEN-19, "Inter-Network Naming,
Addressing,

> and Routing"; although it is very old (and thus somewhat out of date), it
> is
> still full of useful insight on these ideas. RFC-1498, "On the Naming and
> Binding of Network Destinations", is also very important.
>

Thank you for the reference. Only that who wrote it, and whether I buy that
concept.

-- 
Regards,

DY
http://cnu.kr/~dykim