[OSGeo-Discuss] standard for pushing data from GPS device to a web service

[Mateusz Loskot]

Anton Martchukov wrote:
> On Thu, Jan 18, 2007 at 09:06:53PM -0500, Allan Doyle wrote:
>>>> I think using a globally unique id to identify each gps unit would be
>>>> fine.
>>> Rich,
>>>
>>> Do you know any such GUID defined for GPS units?
>> The bluetooth MAC address will be unique. Not sure what to do about  
>> serial port GPS's. Maybe use something like your domain name? Java  
>> style - net.loskot.mateusz.gps
> 
> Basically, since no existing protocol was suggested and so
> we will be stuck with developing our own one for interaction
> between network-enabled GPS unit and the server, we will use
> SOAP and HTTP authentication scheme.

I think WFS(-T) was a kind of suggestion.

> In HTTP authentication we can use text string as login that
> is some unique GPS id (at least in the scope of given
> server).

Yes, if particular application needs to identify a device, then it could
be solved by application-specific identifier system,
with some help on the protocol side.

For example, user/application is configured to authorize with remote
service using ie. e-mail address + password, transmitted by a WFS-like
custom protocol.

> Since it looks like no global GPS unit MAC/id already exists,
> we need some authority that can control the process of
> assignment of such ids if we want them to be globally
> unique. 

I believe application user can state such authority or provider of
remote service.

> Our another way it's possible to adopt some existing
> identifier authority for those purposes. Application
> enterprise numbers assigned by IANA for purposes of LDAP,
> SNMP and maybe some other protocols comes to mind. I think
> that it should be possible to use IANA assigned numbers for
> this. Everyone, who want to develop such network-enabled GPS
> unit, will have to request a number in IANA and then assign
> numbers for each unit within it's own namespace he got from
> IANA. That will work just like MAC addresses work.

Yes, but first simple solution could be implemented and tested, then we
can see if it's enough or not.

I mean, there are many Web services like YouTube, Yahoo!, Google, etc.
every service provides it's own authorization system, based on HTTP(S)
protocol. Similarly, every provider of remote service for GPS users
should has a liberty for implementing its own authorization system.

> But, can someone advocate the necessity of using globally
> unique ids for units? 

I can't and I don't think it's needed.
I just discussed what is available and what is not, in this matter.

Cheers
-- 
Mateusz Loskot
http://mateusz.loskot.net