Terminology [Re: Some Comments on ID/Loc Separation Proposals]

[Brian E Carpenter <brc@zurich.ibm.com>]

After some travel and some flu, I've caught up with this thread, but I 
think it's clearer to comment on Dave's original post.

Firstly, co-chair hat on, we need consistent terminology. When we use a
word, we should all try to use it in the same sense. In the end, we may need
a specific multi6 terminology draft. It is also best if we use existing
definitions rather than inventing our own with subtle changes. So, for example,
if we do use "stack" let's agree on the NSRG definition unless we have a
very strong reason to change it. Otherwise, confusion lies ahead.

Co-chair hat off. Personal comments below.

Dave Crocker wrote:
> 
> Folks,
> 
> draft-crocker-mast-analysis (which will be
> draft-crocker-multiaddr-analysis in its next version) has a section
> for defining some terms.  The section has been prompted by exactly the
> kind of vagueness that the current thread is also trying to fix.
> 
> My primary interest is in having precise definitions that we all find
> useful and use consistently. Some of the current versions of
> definitions in the draft are:
> 
>    Endpoint
> 
>            refers to "the fundamental entity of and end-end
>            communication" [EID]. It is an end-system that participates
>            in an association. Endpoints are distinguished from
>            intermediate, infrastructure nodes and from hosts.

Yes. But this is too generic to be really very useful... and trying to
get more precision always ends in a rat hole.

> 
>    Identifier
> 
>            refers to a unique label for an endpoint. The label is used
>            simply for distinguishing one endpoint from another.

But to use OSI terminology, this doesn't tell me enough, because I don't know 
whether you are talking about a layer 2, 3, 4 or higher endpoint. That's the
advantage of "stack" - we know it's layer 3.

>            Because a locator is usually globally unique, it might be

I wish. But every single time I travel, I find myself ending up with an
identifier that *isn't* globally unique. I don't think "usually" is at all
accurate. "supposedly unique" might be more accurate.

>            able to serve as an identifier. However this use will often
>            suffer administrative and referential limitations as a
>            global identifier for mobile endpoints. This is exemplified
>            by the current problems experienced with the dual role of
>            IP Addresses.

I don't think so. You're correct of course that id/loc overlap is a problem
with mobility. But the problems *today* are caused mainly by non-uniqueness
and that is what is exemplified each time you hook up in a hotel.

> 
> Suggestions for improving the text are eagerly sought.
> 
> As with others, I do not think it is useful to have ID refer to an
> interface.  Stack, endpoint or process all seem more helpful.

True. Interfaces are a somewhat elastic concept and not interesting
at systems level. Personally I prefer NSRG-stack to endpoint, because
as mentioned above it has more precision.

> 
> Speaking of 'stack', what definition text would folks like.  The NSRG
> paper introduces the construct nicely, but I'm not sure the text there
> is what we should live with.

Do you have precise criticisms of that definition?

> 
> For that matter, what is the difference between endpoint and stack?

As an ex-NSRG member, I can tell you that "stack" was chosen because
people couldn't reach precise consensus on what "end point" means. 

   Brian