RE: crypto delay (Notes from Non-meeting)

> Now on to other things..
>
> What is the location object be are all talking about ? XML?
> ASN.1 <just
> kidding>
Oh, I think XML is pretty good. Generally, I think we plan to copy
something that works, although we have some issues others may not have
to deal with.

>
> What is the semantical expression of location object the PSAP needs?
> Typically it keys a database look up based on TN to address.
> PSAP address
> databases are simply subsets of LIDB or DA information
> supplied by the
> LEC's this is now complicated by the mobile environment where formal
> Lattitude and Longitude are used ..again keyed to address.
My thinking is that the object has both lat/lon/altitude as well as
planet/country/state/county/city/street/building/floor/room/cube, with
translation flags. In the emergency call scenario, you start with one,
translate to find the other if you can, look up the lat/lon in the
database to find which PSAP, and send both to the PSAP.

(security note - translators muck with the data; have to be in the
trust loop).

>
> So in those environments where E.164 addressing is uesd that is a
> requirement...not we have the problem of sip URI's
Yeah, for now you actually send the E.164. We may need to create a database
of E164s to PSAP (I hope not - depends on the gateway network architecture)
Note to you privacy folks that in this case E.164 = location - there is
an existing database that translates E.164 to street
address/building/floor/...

>
> Now what other paramaters does it need to transmit?
>
> TTL?
Oh, cute, I like it!

>
> Confidence Factor ? By this I mean ..whatever system generates the
> location information may be able to transmit its Confidence in the
> reliability of the data. This might be useful in those cases
> where L/L
> triangulation is difficult or is unable to reach pinpoint accuracy.
Yeah, the whole object actually needs an accuracy estimate in it as
well as a confidence. The spec needs a way to intentionally degrade
accuracy, but that's another story.

>
> I'm just thinking out loud about what could or should not be
> included in
> the object.
Well, we have discussed previously that you may need a time stamp.
Some folks want motion vectors
There are things associated with the datum used when you have lat/lon
The translation thing is really more than a flag; you probably want
some kind of identifier (signature in some cases) of who did the translation

I'd love to actually start working on that stuff - it's real engineering,
and we can actually get it done in a finite period of time.

Brian
Received on Wed Apr 3 17:03:30 2002