[Asterisk-Users] New ENUM service, what do you think?

[John Todd]

At 7:14 PM +1000 on 5/4/04, Duane wrote:
>John Todd wrote:
>
>>TRIP (RFC 3219) is the answer, but I'm the only one pounding that 
>>drum, it seems.  If anyone here on the list has $100,000 to put 
>>together a real programming effort towards getting that 
>>implemented, y'all let me know.  The longer this waits, the more 
>>lame and broken become the solutions offered.  <sigh>
>
>One small oversight in your thinking, something like TRIP will only 
>benefit large telcos and VOIP providers with interconnects, I don't 
>see this flowing down to a tangible benefit to the average person, 
>where as something like enum.164 is.
>
>TRIP is based on BGP and BGP already does most of the IP routing 
>smarts TRIP is supposed to be beneficial for, however that $100k 
>would be better spent on improving the smarts in the call routing 
>software rather then turning things back into a hub and spoke model, 
>p2p is way more efficient if it can be utilised to it's full 
>potential.
>
>At this stage the only potential method to prevent VOIP spam is 
>something like SPF records, which would only end up duplicate enum. 
>It's a lot harder to get phone numbers then IP addresses, so this 
>would overcome people's concerns about dynamically allocated IPs, 
>phone numbers aren't.
>
>--
>Best regards,
>  Duane


I strongly disagree with your summary that TRIP doesn't help the 
smaller user.  In fact, the reason I'm so strongly an advocate of 
some type of TRIP development is that it removes the barriers for 
small entities in the pursuit of better call rates for TDM offload 
and VoIP interconnection.  Comparative routing data should not be the 
sole domain of huge telephony firms.

One example...

Currently, I see quite a few people here trying to get good rates to 
various international destinations (regardless of their nation of 
origin.)  Wouldn't it be nice to have a protocol that allowed the 
home or small business user to have COMPETING long distance carriers 
on a per-call basis?  When one of them runs a sale, your voice 
traffic could (according to your rules) shift over to the least 
expensive/best sounding/whatever carrier that you'd chosen.  Just get 
a TRIP feed from three or four carriers, and away you go.  It all 
would happen automatically, and you could preference or de-preference 
certain metrics as you went along but the carriers will be sending 
you their most up-to-date routing information for PSTN handoff 
destinations.  Wouldn't it be great if your Asterisk server had that 
ability?  This is just one use and benefit case of TRIP; there are 
many others.

If you say that ENUM is going to solve that problem by offering 
pointers for every phone prefix in the world in the next 5 years, or 
even 33% of them, I would suggest that is a rather optimistic 
outlook.  ENUM cannot have competing answers to the same question; it 
MUST have a single answer, no matter how many private ENUM servers 
you put in the path (otherwise, you're just redesigning TRIP.)  TDM 
offload in between VoIP networks is here to stay; we just need a 
protocol that allows inter-system route exchange for those of us 
lucky enough to be able to take advantage of it today, not sometime 
in the far off future.  Yes, it will also help large carriers as well 
for their exchange of route information, but it's not limited to 
their use.

TRIP is like BGP in it's design, but extremely different in it's 
implementation.  It layers on top of IP, so arguments comparing BGP 
to TRIP with terms like "hub and spoke" are invalid.  Destination 
information does not (necessarily) follow any of the path of the 
lower layers of the routing protocol.  Additionally, I am unclear on 
how you believe that TRIP is involved in "IP routing smarts."  The 
two are not linked in any way.  Can you clarify?

I am uncertain to what your final comments about spam refer.  Neither 
ENUM nor TRIP address issues of call validation in a realistic 
manner; any SPF-like methods for verifying origination work equally 
well with either reference scheme.  Remember that ENUM is a stopgap, 
and we should do all we can to move away from numbers as an 
addressing scheme for VoIP (or any protocol) delivery.  My SIP phone 
address is "jtodd at loligo.com" but the only reason most people can't 
use that is because they are crippled by phones with numeric keypads. 
ENUM is the in-between method to map numbers to more flexible 
addressing until we have smarter phones on our desks and we can use 
the more flexible addressing methods to "dial" the other party.

As I've said, I am a firm believer in ENUM as a "second-generation" 
VoIP routing method, but I'm just as firm a believer (due to very 
hard-won experience in the PBX and carrier markets) that it is 
insufficient at this time to make any difference at all in anything 
other than the most theoretical environments, or environments that 
have been jury-rigged to use ENUM because there was nothing better 
available.

JT