On Fri, 29 Jul 2022, 16:04 Massimo Candela, <massimo@ntt.net> wrote:


I'm not against making the purpose of the database more explicit towards
geolocation data. I was just commenting that I don't see why the current
generic purposes are not good enough (especially since we have a bunch
of other attributes, including "geoloc").

My argument is that none of that bunch of attributes are explicitly covered by the current purposes in a clear and obvious way. 



On 29/07/2022 11:22, denis walker wrote:

>>
>> Geofeed wants to correct geolocation problems. The geofeed attribute is
>> exactly a way to "coordinate between network operators" (with, or
>> without, the intermediation of geolocation providers). Geolocation
>> problems impact the availability/performance of content/services. The
>> medium is the network. Geolocation problems are network problems.
>
> No, "geofeed:" is not covered by this purpose. It may be provided by
> network operators but it is not 'used' by them.

And by who is used?

I think I've been suitably corrected on this point :)

Cheers
denis 
Co- chair DB-WG 


> It has nothing to do
> with 'network problem resolution

This is a weird statement. The network has a physical/geographical
component, is not just topology and protocols.

We have CDNs doing great moneys by offering solutions to this problem.

> It is, as
> I suggested, data that is used by external services.

The data is not -used- by external services!

There are external services, like geolocation providers, aggregating it,
but the data is -used- by other network operators on the other side.

Even more, some operators started collecting geofeed data directly,
without the intermediation of geolocation providers. A great example of
it is Google.

The producer and the final user is always a network operator.

ISP -> content
content -> CDN
content -> transit
ISP -> geoloc provider -> content
etc.


Ciao,
Massimo