Re: Augmenting mobile-broadband-provider-info

["Mark Haack \(werk55.de\)" <Mark Haack werk55 de>]

Hi
----- Original Message ----- 
From: "Dan Williams" <dcbw redhat com>
To: "Mark Haack" <Mark Haack werk55 de>
Cc: "Marcel Holtmann" <marcel holtmann org>; "Antti Kaijanmäki" 
<antti kaijanmaki nomovok com>; <networkmanager-list gnome org>
Sent: Wednesday, January 20, 2010 10:15 PM
Subject: Re: Augmenting mobile-broadband-provider-info


On Mon, 2009-11-23 at 12:15 +0100, Mark Haack wrote:
> Hi
> ----- Original Message ----- 
> From: "Marcel Holtmann" <marcel holtmann org>
> To: "Antti Kaijanmäki" <antti kaijanmaki nomovok com>
> Cc: <networkmanager-list gnome org>
> Sent: Monday, November 23, 2009 10:30 AM
> Subject: Re: Augmenting mobile-broadband-provider-info
>
>
> > Hi Antti,
> >
> >> > > Do you mean the international dial code like +1 for the US, etc? 
> >> > > We
> >> > > could do this.
> >> >
> >> > Yes, that's what I mean.
> >> >
> >> > > At least in the case of the timezone db, I think there's better
> >> > > places
> >> > > to put that.  I realize not all platforms use glibc, but I have to
> >> > > believe that even if you don't, there's going to be some
> >> > > timezone/locale
> >> > > information already on the system that it would be a shame to
> >> > > duplicate.
> >> >
> >> > Ok, I agree with that. Could we at least add countries and provider's
> >> > that do not necessarily offer mobile broadband but just telephony? A
> >> > mapping from MCC MNC to country and provider name can be helpful in
> >> > various ways.

Please pay attention to it, because country name by standards only fits 98% 
of the real world.
I.E. you have this little isles near UK or think of Palestine Networks.

Sometimes you also need extra geo information to display i.E. india with all 
of its regions.

for quick reference I usually use
http://www.numberingplans.com/?page=plans&sub=imsinr


> >>
> >> I agree. This would turn m-b-p-i into cellular-provider-info.
> >>
> >>
> >> When we are now talking about augmenting, I would like to make some
> >> proposals also. I would like to have optional network-name field added
> >> to gsm section. Virtual providers use same parent networks with same
> >> mcc/mnc pairs and there's no way telling them a part. Here's an example
> >> with two Finnish operators which both operate on DNA network:
> >>
> >> DNA:
> >> at+cops=0,2
> >> at+cops?
> >> +COPS: 0,2,"24412",2
> >>
> >> GoMobile:
> >> at+cops=0,2
> >> at+cops?
> >> +COPS: 0,2,"24412",2
> >>
> >> Fortunately each provider conveniently happens to claim they "own" the
> >> network they are operating in:
> >>
> >> DNA:
> >> at+cops=0,0
> >> at+cops?
> >> +COPS: 0,0,"dna",2
> >>
> >> GoMobile:
> >> at+cops=0,0
> >> at+cops?
> >> +COPS: 0,0,"go.mobile",2
> >>
> >> We can use this long alphabetical format of the network to identify the
> >> virtual providers from each other:
> >>
> >>         <provider>
> >>                 <name>Dna</name>
> >>                 <gsm>
> >>                         <network-name>dna</network-name>
> >>                         <network-id mcc="244" mnc="12"/>
> >>
> >>                         <apn value="internet">
> >>                                 <dns>217.78.192.22</dns>
> >>                                 <dns>217.78.192.78</dns>
> >>                         </apn>
> >>                 </gsm>
> >>         </provider>
> >>         <provider>
> >>                 <name>GoMobile</name>
> >>                 <gsm>
> >>                         <network-name>go.mobile</network-name>
> >>                         <network-id mcc="244" mnc="12"/>
> >>
> >>                         <apn value="internet.gomobile.fi">
> >>                                 <dns>217.78.192.22</dns>
> >>                                 <dns>217.78.192.78</dns>
> >>                         </apn>
> >>                 </gsm>
> >>         </provider>
> >>
> >> This allows us to automatically select the correct provider.
> >>
> >> Any thoughts?
> >
> > you can't trust the network name string returned by AT+COPS since there
> > are so many factors coming into play here. So first of all you have the
> > names stored in the modem itself, then the names stored on the SIM card
> > and then the potential updates over the network. Every hardware does
> > different things to present the result of AT+COPS.
> >
> > And to make it even more complex, in case of roaming situations some
> > devices actually to weird concat of home network and current network.
> >
>
> To blow on the fire, in  (if I remember right) i.e. US Cingular there is a
> technical roaming without actual charge, because the subnets belong to 
> each
> other. Which means on technical level there is roaming, which is reported 
> by
> +creg, but it is not the right business logic, which belongs to the user
> domain.

> T-Mobile US has completely free GPRS/EDGE roaming for their prepaid
> customers throughout the US, on a variety of providers.  Like you say,
> the roaming logic is really in the billing systems and that's not
> something that's easy to track.  AT&T phones will also spoof the
> provider name shown on the phone and always show AT&T even if you're
> roaming on some other provider.

I guess it is still possible to track business logic.

1. We can identify SIM Card
2. Even if PLMN is spoofed it is AT&T so we know it is free for prepaid 
customer anyway

The main challenge is to put all the logic into code, db or whatever and 
maintenance of it would be very weird.

BR
Mark


> Dan