Re: Terminology discussion

["Leon Gommans" <lgommans@xxxxxxxxxxxxxx>]

Freek,

We started out the discussion with the term "GLIF resource".
Let's go back to this term for a moment.

The term "GLIF resource" can be considered from an almost infinite
amount of viewpoints. This fact, imho, causes a lot of confusion.
As this is a human thing, everybody will use his or her own set of
viewpoints to consider a definition of a term.

I made the remark during the discussion to add the term "physical"
in an effort to confine the definion towards a tangible entity
that interconnects two points. What a realy wanted  to say is:
consider the GLIF resource definition from a particular viewpoint.
The additional term "physical" may refer to an "engineering"
viewpoint which, focuses on the deployment aspects.
E.g. a physical GLIF resource may be deployed between two GOLE's.

One may also consider the "technology" viewpoint where you focus on
specific technologies, both hardware and software, which will be
used to implement a  GLIF resource (SONET/WDM/DWDM/Ethernet
UCLP/DRAC/GMPLS etc etc.).

Both viewpoints are sort of borrowed from the ODP model. The ODP
model also considers the "enterprise" viewpoint, from which I loosely
derive below viewpoints:

One consider a GLIF resource from the economic viewpoint. Here it
would point at the ablility to obtain the use of a link as an
economicaly scarse good

One could look at a GLIF resource from the legal viewpoint, where a
legal entity can make rules that govern the usage.

Other viewpoints may be administrative, functional, security etc. etc.

Important to note is that each viewpoint uses a viewpoint language to
develop a viewpoint specification.

IMHO it would be useful to first identify a number of viewpoints
which we think are useful and pragmatic to consider one or more
defintions of a GLIF resource.

Regards .. Leon.

> Just yet, there was a plenary discussion on some basic terminology for
> GLIF:
>
> * GLIF physical resource:
>    - Lambda
>    - GLIF Open Lightpath Exchange
> * Physical resource owners
> * LightPaths (and services)
>
> Allow me to stir up the discussion a bit more, but this time on the list
> instead of plenary.
>
> I think it might not have been clear that (in my humble opinion) this is
> a not a technical defination. For technical definiation, we really need
> more detail.
>
> Let's give an example about "resource owners". It can mean "legal
> owner", "economic owner", "administrator", "operator". (e.g. IEEAF may
> pay -economic owner- for a Global Crossing -legal owner- circuit,
> administrated by SURFnet, and operated by SARA).
>
> However, this is not so convienant for our lunch-time chit-chat (or ,
> unless you're really a TMN geek, so we need some more basic terminology,
>   which we should be able to explain to our mother (unless you're Lars
> Fischer from NorduNet, since he just confessed his mom DOES have a
> telecom-background).
> Such a basic terminology should be terminology independant, or we end up
> changing it every few year. If you're really interested in that, please
> read ITU G.805, or <shameless plug>watch our work on network description
> languages</shameless plug>.
>
> Basically, there are tree technicall elements, which have been defined a
> lot of time in different protocol.
>
> * Central locations
> * Connection between those locations
> * An end-to-end connection
>
> Locations are hops or exchanges, and may be complex, have service, or
> even be distributed if that's what we want.
> However, we don't define the details here yet. Just the basic terms.
> A connection is a basic, transport capacity between two locations.
> These basic connections can be divided into channels, to allow multiple
> datastreams on the same connections, and then concatenating those
> channels (or connections) into a tandem which results in an end-to-end
> connection for use by end-users or applications.
>
>
> In Graph theory, this is respectively:
> vertices, edges and path
>
> In ITU G.805 it's called:
> connection point (CP), link connection (LC) and network connection (NC)
>
> In telecommunications, it's called:
> exchanges (or switches), trunks and circuits
>
> In IP, it's called:
> hop, a Layer 2 link (or link-local connection), connections
>
> In my day-to-day talk, I call this:
> transport exchange, link and path
>
> Someone else may call it:
> open optical exchange, transport service and e2e connection
>
> Well, and in GLIF it's now (propesedly) called:
> GLIF Open LightPath Exchange (GOLE), lambda and LightPath
>
> Yes, I do think these particular words sucks. It's loosy. (I would
> obviously prefer my words), but at least it avoids confusion, so I'm
> fine using these words in this community.
>
> Regards,
> Freek Dijkstra
>
>