Ahh a memory I'd rather forget, unknown to most, John Todd and myself started a free enum service, similar to what you're doing. (it was called freenum.org) Unforunately, the project never really got going, due to lack of time and interest (after thinking it over). I believe it would never have
A worthy cause, to be sure, but there are many worthy causes out there, all of which take 8 hours a day. :-)
enough numbers to warrant an enum lookup every time you call. Wasting time doing a dns lookup for the 1 in 1000 chance an enum entry will be there isn't worth it. Also, as soon as you get forged numbers (someone taking over other people's numbers), you'll be in big trouble. "Our number was taken over and we lost 20% of our business as a result, i want a million dollars"
My goal for the project was to have it housed/resolved from servers located in a country that did not have overly aggressive corporate legal rights. None of the northern Europeans I spoke with had the time to work on such a development from the legal perspective, and I don't have any other contacts in such nations.
Here's a few points:
1) authenicating numbers - JT correctly pointed out, you can't allow people to call you to verify as caller id can be spoofed. He proposed a group of asterisk servers calling for verification. I was going to write into this advertising info so you could get businesses to do the calling for you eg "you will be contact by an * server, sponsored by blah <insert small banner or link>"
I think that this would be a minor cost issue for someone who wanted some good press for their VoIP service, based on the number of calls and the time curves. International might be a struggle, but perhaps gateways in commonly-accessed nations could be obtained.
2) DNS - IMO, bind just won't work - PowerDNS or similar I'd suggest, dumping a zone file from mysql when you reach large numbers of entries doesn't scale
3) You need to work out a good and easy way to verify companies (ranges of numbers). Targetting the single line people I don't think will yield you enough numbers.
Yes, this is true. I came up with the (bad) idea that perhaps the first and last number in a range would be targetted, and then some small (1%?) of numbers in the middle of the range would be pre-chosen and the submitter would be told well in advance when those numbers would be tested, such that they would have someone answering on those lines. Limit the blocksize to something reasonable (1000 numbers? 500 numbers?) so that spoofing would be kept to a minimum.
4) I think you need to allow users to either point their entry to their DNS or make an easy interface that will generate an entry for them. Don't force them to enter raw E164 entries (but let them if they really want to)
5) make a non profit organisation, or you'll get sued personally.
Indeed, this is critical, or see my point above about "neutral nations."
good luck, I'm sure JT will have a few comments (probably cursing my name)
No, not at all. I too, have many projects to do right now as it stands, and while I think that an "open" parallel ENUM root is an excellent effort, I also have to keep perspective on the other projects in the works. ENUM is easy to put off, since we can all see real ENUM root service "just over the next corner..." <cough, cough>
JT
Adam
Anyone wishing to help build/manage openenum.net please contact me via email [EMAIL PROTECTED] ... I would like to have someone assist in building and management.
Thanks, bkw
_______________________________________________ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users
_______________________________________________ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users
