Re: [Geopriv] Fwd: I-D Action:draft-ietf-geopriv-arch-00.txt

["James M. Polk" <[e-mail]>]

At 01:53 PM 7/14/2009, Alissa Cooper wrote:
>James, Ray,
>
>The last time we were discussing lo-sec, it was obvious that some
>clarity was needed about a single entity performing multiple Geopriv
>roles.

I agree that a single entity can perform multiple roles - but that's 
the opposite of what you're suggesting here; as you're changing the 
same function and definition of the entity based on who's asking for 
the same information -- which is what I object to.

>Throughout the new version of the document, we've tried to make
>that clear, both by stating it explicitly ("It is not uncommon for the
>same entity to perform both the LG and LS roles, or both the LR and LS
>roles." in section 2), and by consistently referring to "the entity
>acting as L_" or the "the entity performing the role of L_".

again - that's the opposite of what's happening here.

What you are saying is that a LIS is the dereference server and 
destination of a location URI from the target *only*.

If a 3rd party wants that same target's PIDF-LO using the _same_ 
location URI it is no longer contacting the LIS, it is contacting a 
Location Server.

This is nuts.

LISs should be used for 1 purpose, as should Location Servers.

A LIS should be used to provide LI to a Target via an LCP.

Location Servers are location inserters (i.e., the SIP UA or Proxy 
that inserts location into the SIP request).

We need to have another singular purpose term defined as the 
destination of the location URI, call it a Dereference Server (DS) 
which provides PIDF-LOs from Location Recipients that were given a 
location URI from or of a target.

This means everything is clearly singular in


>This holds true for LIS as much as any other Geopriv role. Nothing
>prevents an entity acting as a LIS from also acting as an LS. Perhaps
>my one-sentence description was more confusing/objectionable than the
>text itself. From section 3.2.2:
>
>>Some performing the Location Generator role are designed only to
>>provide Targets with their own locations (as opposed to distributing
>>a Target's location to others). The process of providing a Target
>>with its own location is known within Geopriv as Location
>>Configuration, and the LG that provides such location is often
>>described as a Location Information Server (LIS). Protocols used
>>exclusively to communicate location from a LIS to a Target are known
>>as Location Configuration Protocols [8]. Several such protocols have
>>been developed within Geopriv [9][10][11][12].
>>  By definition, a LIS needs only to follow a simple privacy- 
>> preserving policy: transmit a Target's location only to the Target
>>itself. This is known as the "LCP policy."
>>Importantly, if an LS is also serving in the role of LG and it has
>>not been provisioned with Privacy Rules for a particular Target, it
>>MUST follow the LCP policy, whether it is a LIS or not. In the
>>positioning phase, an entity serving the roles of both LG and LS
>>that has not received Privacy Rules must follow this policy. The
>>same is true for any LS in the distribution phase.

>The notion of a LIS in the role that provides a Target with its own
>location seems to be supported in the rest of our existing geopriv
>documents (lis-discovery, lbyr-requirements, held, l7-lcp-ps).

"...the rest of our existing geopriv documents..."

These are all HELD originated and HELD only documents.  Surely 
Geopriv has produced more documents that just about HELD, hasn't it?

If so, then this quote isn't so accurate, is it?

BTW - which of these documents define how HELD returns a PIDF-LO in a response?

>Can you  provide references that instead point to LIS as a 
>dereference server
>that responds to any LR's request?

Can you?

A LIS has always been our least defined entity.

As suggested above, why get into a discussion about an acronym (LIS) 
that does a function that accurately describes what we're talking 
about just fine (i.e., Dereference Server (DS)).

This would solve all our problems and minimize or eliminate 
everyone's confusion.

>Also, we will get authors/acknowledgements squared away in 
>a  subsequent draft.

yeah -- and the 2 chairs are part of that author team... so who's 
running this document (as a matter of process)?

>Alissa
>On Jul 14, 2009, at 4:35 AM, [e-mail] wrote:
>
>>
>> > >* Location configuration is called out explicitly. LIS is defined
>>as a
>> > >special case of LG that only provides a Target with its own
>>location.
>> >
>> > err, this is yet another definition change in the grand scheme of
>>Geopriv...
>> >
>> > so, a LIS is an LCP server, and what is the destination of
>> > subscriptions for dereferences - but only by the target itself, not
>> > dereferences by other entities?
>>
>>I'm not happy with this either.
>>
>>What are the motives for this change?  It seems to me that
>>generically a "LIS" is widely taken to mean anything that implements
>>HELD (or similar?) regardless of who is requesting the location and
>>to restrict the terminology now seems rather odd.
>>
>>Ray
>>
>>--
>>Ray Bellis, MA(Oxon) MIET
>>Senior Researcher in Advanced Projects, Nominet
>>e: [e-mail], t: [phone]
>>
>>
>>_______________________________________________
>>Geopriv mailing list
>>[e-mail]
>>https://www.ietf.org/mailman/listinfo/geopriv
>
>
>
>
>
>
>
>