As someone who participated in the LoST development, things like this always 
look complex until you try to build something "simpler".  It's just not a 
simple problem.  You can make it simpler only by changing the problem.  For 
example, you can require some amount of provisioning before it will work.

The way LoST works, the national regulator can control how resolution of the 
LoST servers works.  You can have all queries within a country go to a specific 
server that then refers or iterates to another server, or you can return a list 
of servers that can satisfy the response and let the client choose.

Brian


On Feb 18, 2013, at 10:52 PM, Weixinpeng 
<[email protected]<mailto:[email protected]>> wrote:

Thanks for your comments. Please see my answers inline.
Best Regards,
Xinpeng W.

From: Peter Stanforth 
[mailto:[email protected]<http://spectrumbridge.com>]
Sent: Tuesday, February 19, 2013 2:09 AM
To: Weixinpeng; [email protected]<mailto:[email protected]>
Cc: Peter McCann
Subject: Re: [paws] [A new PAWS DB Discovery Draft]

I am Still not a fan of basing DB discovery on the LOST protocol. It seems to 
be way to complicated and I would hope someone can explain how to reconcile the 
capability with the need.
That said the following comments are independent of the LOST protocol.

The introduction should be expanded to identify that there are many ways that a 
device could get access to a database not just the manually programmed example 
given.
[Wei] Yeah, you are right here, I there are many ways for a device getting 
access to a database besides manually programmed, and I will describe more in 
detail in the later version.

The WSDB DS will very likely be nested. I cannot see Ofcom giving up ownership 
and authority of its DB discovery process for instance. So a WSDB DS would have 
to reconcile that a WSD was in the UK regulatory domain and provide a reference 
to the Ofcom DB Management system.

Under current US rules there is no concept of a WSDB DS radios are certified 
against a database. However a device certified in the USA may travel to other 
countries and then need to use a WSDB DS. So the other case that needs to be 
discussed is a device that in some instances has a preconfigured link to a 
database and the ability to use a WSDB DS in other cases.  There may be a 
situation (and I am making this up for an example) where a device is certified 
in the US along with a database, certified in Canada along with a database and 
certified generically in the rest of the world. This device may have a 
hierarchy where it checks with the US DB first. If it gets denied (because it 
is outside the regulatory domain) then it tries Canada and if that also fails 
goes to the WSDB DS.

This raises a couple of questions. Is the WDSB DS going to return all 
"certified" databases in a regulatory domain or only the ones that are 
appropriate for the device making the query?
[Wei] Here the databases return from WSDB DS means all the certified databases 
in a regulatory domain. Whether a database is appropriate is not decided in the 
discovery procedure, but in the following procedure when WSD connects to the 
database, if the selected database cannot serve the WSD, the WSD will ignore 
the database and try another one and so on.

How is the set of regulatory certified domains mapped to those databases 
willing to serve the device or the databases the device is willing to go to 
(This is a business decision)? Wearing my DB provider hat, I only want to (and 
only will) serve devices for which I have a business relationship with the 
manufacturer or possibly the owner.
Does this process work if the owner of the device is to make the decision 
rather than the device?
Did we cover the correct set of responses from a DB to cover this mechanism?
       Although the WSDB DS thinks that the device should be associated with a 
DB the DB does not
       The DB does not want to serve the device even though it is on the WSDB 
DS list
       The Device does not want to work with a DB on the list (how does it know 
if all it gets is a URL – so radio vendor x is won't work with  DB provider y)
       The device has a preference to work with DB vendor y in as many 
regulatory domains as possible

It seems many of these can be addressed more easily if the device uses a 
manufacturer provided discovery process (as it will have knowledge of the 
business relationships) rather than some new global entity that will have to be 
created, maintained and supported.  Which brings me to my final question who 
owns and operates the WSDB DS and who pays for it?
[Wei] The discovery mechanism provided here is an optional method and it may be 
cooperated with other method to find the appropriate database for WSD.

Regards,
Peter S.


From: Weixinpeng <[email protected]<mailto:[email protected]>>
Date: Sunday, February 17, 2013 9:20 PM
To: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>>
Cc: Peter McCann <[email protected]<mailto:[email protected]>>
Subject: [paws] [A new PAWS DB Discovery Draft]

Hi all,
A new draft of DB Discovery has been submitted:
http://www.ietf.org/id/draft-wei-paws-database-discovery-00.txt

Comments are welcome!!

--Xinpeng Wei.
_______________________________________________
paws mailing list
[email protected]<mailto:[email protected]>
https://www.ietf.org/mailman/listinfo/paws

_______________________________________________
paws mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/paws

Reply via email to