Re: [Enum] I-D Action: New draft - draft-bellis-enum-send-n-00

[Lawrence Conroy <lconroy at insensate.co.uk>]

Hi Clive, Penn, Ray, folks,

I am somewhat discomforted by this thread. IMHO it's a damn good idea  
having (in effect) an digit length analysis tree in DNS. It sure as  
heck beats looking through the different CSV/Excel/Access/... formats  
in which different Regulators publish their number plans.

However, the ENUM standard doesn't seem to be set up to allow  
definition of these useful records. This may not be appropriate for  
E2U, but instead be another DDDS application entirely. Otmar, Alex,  
and Michael have had similar ideas, and they didn't prevail even  
though they were neat proposals. Heck, Rchard Stastny and I had some  
fun with VOID, and were reminded that THAT won't fly in ENUM with  
block numbers.


Scott has released a draft of rfc3761bis-03.
NOTE WELL - this has NOT changed the applicability text of 3761 at all.
Dialing Plans are specifically excluded from ENUM. Thus before the  
esteemed SIP aficionados dive in, ENUM is about numbers, not dialed  
digit strings.

An ENUM client is also expected to make an ENUM query only where it  
believes it has a complete E.164 number. RFC 3761 goes on to accept  
that a client may not know what constitutes a valid E.164 number (some  
of us call outside the NANP :).

In principle only having a partial view of global number plans is not  
good enough. The idea of provisioning every nation state's number plan  
locally into every switch is surely not what was meant. **

However, it is pushing this a bit far to send out an ENUM query with a  
partially collected number, in the expectation of getting back an ISUP  
SND style response. This is not something the originating node  
believes is a complete number - it's something it suspects isn't. Thus  
whilst I really like the concept, I fear that ENUM is not the  
appropriate carrier.

all the best,
   Lawrence
---------
**
We have had this number plan discussion many times before. In the  
NANP, numbering is (usually) easy. Dialing is another matter, but  
that's way out of scope.
In the UK one can at least know what constitutes a valid number  
length, BUT this is not immediate - one needs to have some digits to  
do the analysis.
Thus whilst a (semi) static analysis tree is possible (see <http://www.ofcom.org.uk/static/numbering/index.htm 
 >), as Clive states, the length of numbers in the UK is prefix  
dependent. Also as Clive points out, there are no subsets - again,  
look at the numbering plan files.
In Germany and Austria, all bets on number length are off, as these  
nation states have open numbering plans. However, again, they don't  
have independent numbers as subsets of others.
This should come as no surprise at all to someone in the business.
I am thus once again puzzled by this resurgence. Is this another game  
of WhackaMole?
----------

On 2 Apr 2008, at 22:03, Clive D.W. Feather wrote:
> PFAUTZ, PENN L, ATTCORP said:
>> I'm having a little trouble understanding the real need for this
>> enumservice. I would think that any network element  receiving a dial
>> string would have its own local information about dial patterns and
>> numbers and so be able to tell when dialing is complete and then be  
>> able
>> to formulate an ENUM query on the appropriate full E.164 number.
>> Since the number length information should be relatively static,  
>> why not
>> just provision it locally?
>
> While the information should be static, it's also very complicated  
> and it
> might not be practical for the network element to store it all. In  
> the UK
> we have number ranges where the length varies in a very localised  
> way; it
> might be that:
>     0456 789 0xx    numbers are all 10 digits
>     0456 789 1xxx   numbers are all 11 digits
> (these are hypothetical examples because I can't get to the relevant
> database right now).
>
> Germany has situations where the customer determines the length of  
> numbers.
>
>> Also, it's not clear to me when the client would initiate the initial
>> query in the example that told it to expect 5-6 more digits.
>
> The client would know, perhaps, that *every* number beginning 0 has at
> least 7 digits. So it would initiate the query after receiving those  
> 7.
>
>> And, if
>> it's 5-6, then if the number has only 5, don't I still have to use a
>> timeout algorithm to determine the user is finished dialing.
>> Am I missing something?
>
> In the UK there are no numbers that are prefixes of other numbers.  
> That is,
> when you've received enough digits, you'll know you've reached the  
> end. But
> you may need several digits to do so - see the above examples.
>
> -- 
> Clive D.W. Feather  | Work:  <clive at demon.net>   | Tel:    +44 20  
> 8495 6138
> Internet Expert     | Home:  <clive at davros.org>  | Fax:    +44 870  
> 051 9937
> Demon Internet      | WWW: http://www.davros.org | Mobile: +44 7973  
> 377646
> THUS plc            |                            |
> _______________________________________________
> enum mailing list
> enum at ietf.org
> https://www.ietf.org/mailman/listinfo/enum

_______________________________________________
enum mailing list
enum at ietf.org
https://www.ietf.org/mailman/listinfo/enum