ENUM Working Group Minutes from RIPE 53

• content to enum-wg-chair@ripe.net
• format to webmaster@ripe.net

Meeting Chair: Niall O'Reilly (UCD - University College Dublin)
Scribe: Alex Le Heux (RIPE NCC)
Jabber: Catherine Carr (RIPE NCC)

Webcast and Feedback Archives:
http://www.ripe.net/ripe/meetings/ripe-53/sessions-archive.html

Start: 14:00

Administrative matters

The Chair welcomes everyone, the scribe and the jabberwock.

There are no additional agenda points to add to the agenda although the order will be different:

http://www.ripe.net/ripe/meetings/ripe-53/agendas/enum.html

There were no comments on the minutes from the RIPE52 meeting, so the minutes are approved.

Review of RIPE52 action items:

52.2 - Carsten Schiefner (Deutsche Telekom)

There has been some discussion with the RIPE NCC. A .zip file has the advantages of making it easy to add more files and it is a single download.

Separate files give a good overview, can be well structured and are the same as the old concept. A downside is that it will produce many hits while searching.

Carsten Schiefner will once more put this matter to the WG Mailing List.

[ ENUM-AP-52.2 remains open ]

The other action items will be dealt with in other agenda items.

Presentation: Operations Update - Brett Carr (RIPE NCC)

http://www.ripe.net/ripe/meetings/ripe-53/presentations/enu-bc-upd.pdf

Questions:

Patrick Faltstrom (Cisco):
Are you doing lameness checks on DNS servers?

Brett Carr (RIPE NCC):
Not yet.

Patrick Faltstrom (Cisco):
Will the RIPE NCC take the initiative for that?

Brett Carr (RIPE NCC):
Yes.

Peter Koch (DENIC):
You had a slide about strange queries, starting with a plus. Have you considered they could be a misimplementation of the DDDS algorithm? Have you looked into from how many different sources these queries originated?

Brett Carr (RIPE NCC):
No, but I have all the stats. But we can. => Action item on the RIPE NCC

[ ENUM-AP-53.1: RIPE NCC to examine and report on "strange queries" ]

Carsten Schiefner (Deutsche Telekom):
The NCC is able to get stats from all servers. Will there be a further effort to break them down by country code?

Brett Carr (RIPE NCC):
We tried, but not all providers have the resources to run the stats on their machines and some are very wary from a privacy point of view.

Otmar Lendl (NIC.AT, via Jabber):
Please get the DNS performance of the Chinese server up to speed (or provide glue for them). Resolving e164-arpa.cnnic.net.cn is quite broken.

Brett Carr (RIPE NCC):
Chinese are working on moving their service to a different server. We are already in contact with them.

Presentation: Quality Task Force - Carsten Schiefner (Deutsche Telekom)

http://www.ripe.net/ripe/meetings/ripe-53/presentations/enu-cs-qtf.pdf

Niall O'Reilly (University College Dublin):
Will you follow up on the Quality Task Force and follow up on the lameness checks by the NCC as an action item?

Carsten Schiefner (Deutsche Telekom):
Yes.

[ ENUM-AP-53.2: Carsten Schiefner to follow up on Quality TF ]

Andrei Robachevsky (RIPE NCC):
Why do you think it's impossible to tell an ENUM Tier-1 registry to be compliant?

Jim Reid:
The answer to that is not technical, it's political. It's run by the country according to that country's rules. So it's a matter of national sovereignty, so I don't think it's the business of the NCC or anybody else to tell them how to run their name servers.

Andrei Robachevsky (RIPE NCC):
Can this issue be brought into the technical plane rather than the political plane? I understand that delegation to a particular entity is a political issue, but running DNS servers is probably more technical.

Jim Reid:
It is a technical issue, but from a government perspective, they feel entitled to do it they way they want to.

Patrik Faltstrom (Cisco):
That's why I asked who took the initiative for doing the checks. There is a very, very big difference between whether a country asks the RIPE NCC directly or indirectly to do these lameness checks and notify them when something is broken and having someone tell a country that they're doing the wrong thing.

Carsten Schiefner (Deutsche Telekom):
The original initiative by the RIPE NCC for the lameness checks was for the reverse mapping space only.

Jim Reid:
It's not an issue of DNS, it's an issue of phone numbers and telephone numbers being an item of national sovereignty.

Daniel Karrenberg (RIPE NCC):
I fully agree with Patrik and what Carsten said. What you want to see is a best common practice written up. That's not telling a country on how they should run their DNS. Being from the NCC I am slightly uncomfortable with the hands off approach that is suggested by Patrik and Jim. In any other constellation I can think of; the parent registry has some rules actually. I'm uneasy about this because the RIPE NCC runs at Tier-0 and any really bad quality falls back on the RIPE NCC. We were getting flak for things not working, so I think the NCC is entitled to make strong suggestions.

Niall O'Reilly (University College Dublin):
(taking chair hat off) There is a delicate balancing act here. If there isn't a safety net that's a bad thing, if the safety net interferes with national sovereignty, that is a bad thing too.

Jim Reid:
I agree with what Daniel says. The RIPE NCC must be careful though with it's role as Tier-0 registry. The instructions to the NCC didn't mention lame delegation checking. If the NCC is perceived to exceed its authority, there could be trouble.

Olaf Kolkman (NLnet Labs):
About the BCP that Daniel mentioned, I think it's more important to provide a set of minimum requirements. Not "this is the bar you should reach" but "this is the bar you should step over" to at least be good.

Field Reports

Presentation - Ondrej Filip (NIC.CZ)

http://www.ripe.net/ripe/meetings/ripe-53/presentations/enu-of-czr.pdf

Carsten Schiefner (Deutsche Telekom):
I would be surprised if a premium rate number holder would actually register it.

Ondrej Filip (NIC.CZ):
It's an option.

Carsten Schiefner (Deutsche Telekom):
If you go commercial in 2007, will you clean the trial database?

Ondrej Filip (NIC.CZ):
No, we will continue on the same platform.

Jim Reid:
What are your plans to incorporate this WG's thoughts on DNS quality and incorporate monitoring into the platform?

Ondrej Filip (NIC.CZ):
No plans.

Sungwoo Shin (KRNIC):
You will use trial user data when you move to the commercial phase. Will there be any validation of that data?

Ondrej Filip (NIC.CZ):
Yes, the trial phase is the same as commercial, except without payment. So there is normal validation.

Sungwoo Shin (KRNIC):
The trial users can use it freely and make a free call without charge?

Ondrej Filip (NIC.CZ):
Yes, you can use it to make a free call.

Presentation - Jim Reid

http://www.ripe.net/ripe/meetings/ripe-53/presentations/ukeg.pdf

Carsten Schiefner (Deutsche Telekom):
Is external DNS monitoring part of the RFP?

Jim Reid:
Yes.

Otmar Lendl (NIC.AT, via Jabber):
Are the SIP URIs in the CRUE records supposed to point to open SIP server, and thus reachable by any caller on the Internet?

Jim Reid:
The SIP server address can point to wherever the carrier chooses, we don't care.

Bernie Hoeneisen (SWITCH, via Jabber):
In case a user overrides his ENUM entry as populated by the carrier: what will appear in the zone file you are directly providing/selling to the carriers? User provided NS RRs or original carrier populated NAPTR RRs?

Jim Reid:
There will be two NAPTR records, if there is a delegation request, they will be replaced with NS records.

Presentation - Sungwoo Shin (KRNIC)

http://www.ripe.net/ripe/meetings/ripe-53/presentations/enu-ss-kor.pdf

Niall O'Reilly (University College Dublin):
Will monitoring be built into the requirements for the service?

Sungwoo Shin (KRNIC):
I don't disagree with Jim's idea.

Presentation - Antoin Verschuren (SIDN)

http://www.ripe.net/ripe/meetings/ripe-53/presentations/enu-av-nl.pdf

Antoin Verschuren (SIDN):
To answer Jim's question: Yes, we will be happy to do so!

Jim Reid:
This is about the consultation you are doing. There are two things: One for redelegation and two, a future meeting where people can comment on the general plan. Is that only for Dutch people?

Antoin Verschuren (SIDN):
It's an open platform.

Presentation - Robert Schischka (NIC.AT):

http://www.ripe.net/ripe/meetings/ripe-53/presentations/enu-rs-aus.pdf

Wolfgang Tremmel (DE-CIX):
Is it possible that different carriers querying get different views?

Robert Schischka (NIC.AT):
No, it's a single public tree. Some carriers might have a different view internally, but not from the outside.

Jim Reid:
Are you using wildcards?

Robert Schischka (NIC.AT):
Yes, we have to, we have an open numbering plan.

Olaf Kolkman (NLnet Labs):
From a DNS client perspective, if I were to query the tree would I need to know the country number plan?

Robert Schischka (NIC.AT):
Yes, you need to know where the 'i' is so you need to know how long the country code is. It's a dirty solution but the only way to go now.

Olaf Kolkman (NLnet Labs):
There may be other alternatives.

Robert Schischka (NIC.AT):
Everybody agrees that in the end it needs to be a separate tree. There was a lengthy discussion in the IETF, the "Dallas treaty" was an interim solution, and will be easy to change.

Peter Koch (DENIC):
This kind of infrastructure stuff is kind of broken. For years we have been hearing that it would take years to convince the ITU. Has there been any effort so far to actually talk to the ITU and start this?

Robert Schischka (NIC.AT):
I don't think the issue has been going on for years, but I don't think anyone is talking to the ITU about this very much. Maybe we should see several countries interested in this, to give more power to the discussion. The tech community wants a neat solution, but the key is buy-in from the telcos, not how the tree is named.

Peter Koch (DENIC):
One of the problems is that the specification is a bit brittle.

Robert Schischka (NIC.AT):
No one in carrier land really cares about this.

Presentation: ENUM Next Generation - Patrik Faltstrom (Cisco)

http://www.ripe.net/ripe/meetings/ripe-53/presentations/enu-pf-nge.pdf

Antoin Verschuren (SIDN):
How are you going to split authority if you do a delegation on for example a path for a number block: if you want to redelegate, for open number plans for instance?

Patrik Faltstrom (Cisco):
There are two possibilities:
1. A neutral party similar to the portability database administrator operates a database.
2. Do delegations higher up in the tree, but those parties must be neutral as well. There is no portability in DNS.

Daniel Karrenberg (RIPE NCC):
How are applications supposed to make use of this?

Patrik Faltstrom (Cisco):
Two ways:
1. Look up the NAPTR records for the phone numbers.
2. If you understand SIP, you can look up _sip._e2u.[phone number] and use that.

Daniel Karrenberg (RIPE NCC): And the _h323 and _message labels are found in some lists?

Patrik Faltstrom (Cisco): In the existing ENUM specification there is a IANA registry for those labels, including a registration mechanism.

Issues for other working groups:

Niall O'Reilly will alert the DB WG and the NCC SERVICES WG as action item 53.3 that we expect to have consensus soon on the NON-REGISTRY issue.

[ ENUM-AP-53.3: Niall O'Reilly to alert other WGs re NON-REGISTRY ]

AOB

The graphic of the T-Shirt will be on Rosie.

Meeting Ends 15:55