[ih] History of Naming on The Internet - is it still relevant?

[John Day]

I knew I would have to wade into this sooner or later.

> On Jul 21, 2025, at 11:46, Patrik Fältström via Internet-history <internet-history at elists.isoc.org> wrote:
> 
> On 21 Jul 2025, at 16:08, Andrew Sullivan via Internet-history wrote:
> 
>> So, to bring this remark back to something to do with history, _why_ did this perceived need arise?
> 
> I find the whole discussion about "naming" has to do with primarily two different questions which are mixed up:
> 
> - Identifier / location split (or mix)

This is a false distinction.  
Saltzer defines "resolve” as in “resolving a name” as “to locate an object in a particular context, given its name.”

IOW, in computing systems one can’t name something without locating and can’t locate it without naming it.

This is why all of the solutions trying to use it don’t scale very well. (I have written a longer essay on this, if anyone interested)

What was trying to be done with loc/id split was to find a work around to the fact that Internet lost the internet layer.
IP was supposed to be in the internet layer, not the network layer.

The early distinction between the Network part and the Host part of an IP address was correct then. Very early on, the host part merely enumerated the hosts within the block of IP addresses. But as the world got more complex, the Host Part came to be assigned in a way that was more compatible with CIDR, i.e., it reflected the structure of the organization’s network to which the block of addresses had been assigned. (Both were locators. This was confirmed by the RRG, when they discovered that the flat EID had to be ‘aggregatable’. IOW , location-dependent. So loc/id split was really loc/loc split.  So by way of analogy to a postal address, the network part or locator part was the ’state/province and country’ (perhaps city) and the Host part was 'the street and building number.’  But that still doesn’t fix that the Internet lost the internet layer.

All addresses name the entity in the layer, that they reside. To use Saltzer again, (N)-layer address is a node address, and (N-1)-layer address is a point of attachment address. (The terms are relative.)
Any point-to-point layer/link does not require an address. (There is only one place for a packet to come out.) It may be assigned a local identifier to distinguish multiple cases of point-to-point interfaces, but a global address is not required. For lower  layers with more than two members, an address is required. It usually a MAC address, but might be a network layer address.

Probably more than 95% of all IPv4 addresses assigned were assigned unnecessarily.

BTW, it has been explained to me that all Internet routers route on the node address. As it was explained to me, "if they didn’t do it, it wouldn’t work.”

This really requires a much longer explanation but this isn’t the place for it.


> 
> - Who decides the lifetime of a "name"

The owner of the name and/or the namespace.

Take care,
John
> 
> Regarding lifetime, I think (influenced by a few discussions with Tim Berners Lee about lifetime of URLs) the most important thing with domain names is that the holder of a domain name can decide what the lifetime of a domain name is, because that is the key thing to a decision about the lifetime of a URL.
> 
> That has for me been a key goal during the years when looking at policies for domain names in ICANN and protocol design in the IETF (epp primarily, but also DNS), the domain holder must be the entity that controls a domain name, and we must minimise the risk of external events that can terminate a domain name.
> 
> Because it boils down to the lifetime of a functioning URL.
> 
> Patrik
> -- 
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history
> -
> Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history