[enum-wg] repost: Proposal for non-geographic ENUM E.164 UPTS for the general public

hi!

>> that's germany. +49 numbers are in their regions fully portable 
>> between pstn providers. currently large regional blocks are 
>> reallocated by voip gateway providers to voip-users. so a voip-user 
>> from e.g. dresden can have a number from e.g. duesseldorf, hamburg, 
>> berlin, munich - or all of them.
> 
> 
> ... and which AFAIK is of some major concern to the regulator. Apart 
> from the fact that they don't have any real means at hand right now, 
> they do not get tired of repeating their unhappiness.

yes - actually (except for some endusers that consider it fun to have 
numbers from different cities) i know of nobody who is really happy with 
this solution, last of all the gateway providers. there's really a need 
for a real solution, and in voip-context a global non-regional number 
space is imho most straightforward. it's hardly optimal to dial 
+4932.... or any other cc-number to reach a user who is currently in 
helsinki and might be in hongkong tomorrow...

as an international prefix could be seen as a public resource that has 
to be available in a non-discriminating way, i beleive personal numbers 
assigned to the endusers are again the most straightforward solution. if 
blocks were allocated to providers and were only portable as the whole 
block, i think that would spoil portability as well as the idea of 
non-regional numbers. actually e.164 provides for other number space to 
be allocated/assigned to companies on a big block level, so when there 
is need for something like that, an existing global upts wouldn't spoil 
anything.

>> that's really ugly but voip providers actually have no big choice. 
>> there is a 'region' reserved for non-regional (in the end of course 
>> still national) numbers (+4932), but this is currently discussed and 
>> it's unlikely that this prefix will be useably anytime soon. also, 
>> there are strong doubts that this prefix will be freely available to 
>> voip-providers, the current discussion shows that there will probably 
>> be dependencies and drawbacks especially for voip.
> 
> 
> That is one of my concerns - that according to the current draft only 
> "real" telcos can get numbers or blocks out of that range. So any VoIP 
> only provider just has to rely on others instead of being able to deal 
> with the regulator directly.

right. although this is a point that shows up quite clearly, the 
impression i get from the current state of discussion is more the one of 
a swamp... it's really hard to get an idea of where it's really going in 
the end. that's really frustrating, while the need for a real solution 
grows every day, and as i said before, i don't think that non-regional 
national numbers are actually a real solution...

> Another one is (and I find that in your proposal, too) the idea of 
> allocating/assigning blocks: at the end of the day we are also talking 
> domains here. In the domain world such a concept is totally unknown (for 
> good reasons!) and would not really be feaseble: get domain names 
> starting with a to k from registrar A and from l to z from registrar B...?!

hm, right, 100% ack.
but maybe that's just my bad explanation in the proposal, the idea was 
to allocate blocks to providers to keep administrative work at the rir 
level low, while every single number stays portable by using the 
hierarchical means provided by whois: if a single number out of the 
provider-allocated block is ported, the maintainer is changed to the new 
provider as well as the enum-delegation (see collection of 'most 
specific' nameserver info in the proposal).
this way every number would stay portable.

hmm - of course, this would spoil 'wish-numbers' (sorry for the brutal 
translation, corrections welcome ;), or even could lead to very very 
strange procedures (i'm too scared to describe what i'm thinking about 
;>) by sbdy trying to get such a specific number...
maybe there is a reasonable solution without allocation of blocks. in 
this case the rirs would probably have to deal with each single number 
request... i'm not sure if that can be automated or handled in a way 
that doesn't produce lots of work at the rir-level... maybe with some 
sort of assignment window? ideas more than welcome! :)

> As you have to ensure portability in the long run anyways: why not 
> assigning a number out of such blocks (whether it's +878xx, +4932 or 
> something else) directly to the user? S/he can port the number to a 
> provider of an own choice without creating "holes" in blocks allocated 
> or assigned to providers/LIRs/...

from my view this would be the most charming solution - but creating a 
lot of work on the rir level is not an option, and currently i don't 
know of a practical solution to that... hmm...

> IMHO the difference really is that people (apart from very few ;-) do 
> not really care about the IP addresses they get because you hardly 
> announce them to third parties as points of contact anyways. But you 
> certainly do with your UPT/... numbers.

right... hmm... maybe without allocation of blocks to providers, but 
allocation of a 'number of (not specific phone-)numbers' could work. 
actually that would already be a kind of AW-solution. hm. sounds 
realistic to me, while work at the rir-level were still rather low.

> That, of course, would be a different registry type - more a domain 
> registry that an RIR. But I am certainly not saying that an RIR cannot 
> or even must not take up that business, too - given that all parties 
> involved agree to it.

i think regarding the whois, it would work.
the most important point is of course the last one, there of course has 
to be substantial consent to a general need for this. from what i see in 
every day business, i'd say it certainly really is there. let's find out...

kind regards,

	Chris Heinze