RIPE 63

• content to [email protected]
• format to [email protected]

Wednesday, 2 November 2011, 14:00
Hilton Vienna am Stadtpark, Vienna, Austria
Co-chairs: Niall O'Reilly, Carsten Schniefner

A: Administrivia (5 min)

Welcome, Scribe, Jabberwok, Microphone etiquette, Agenda

Agenda change: Item E moves after C, before D

B: Minutes of previous meetings

Confirm result of last calls on mailing list (3 min)

Minutes from RIPE 60, RIPE 61 and RIPE 62 are now final. Niall O'Reilly has sent email to the enum-wg list.

C: Review Action List (2 min)

ENUM-AP-61.3 – [DONE] Panel discussion scheduled as item D2

E: ENUM Operations

E1: Tier-0 Report (Wolfgang Nagele, 15 min)

The presentation is available at:
http://ripe63.ripe.net/presentations/121-RIPE63_WolfgangNagele_ENUM_update.pdf

There were no questions.

E2: +46 rescued (Patrik Fältström, 10 min) (via Skype)

The presentation is available at:
http://ripe63.ripe.net/presentations/92-RIPE-ENUM-46.pdf

Carsten Schiefner, DENIC eG, asked if apart from the fact that the delegation had been withdrawn, did the zone technically disappear.

Patrik replied no, the zone was running the entire time.

D: Main presentations (55 min)

D1: Progress with +351 ENUM (Lino Santos, 10 min)

The presentation is available at:
http://ripe63.ripe.net/presentations/90-wg_enum_351.pdf

Carsten Schiefner asked if those participating in their trial also include normal PSTN operators, or only VOIP operators.

Lino replied both.

Carsten asked if compared to the legacy operator, whether they were large or small.

Lino replied that the incumbent is not in the trial and that the second largest operator provides VOIP service at the moment.

D2: Panel discussion on state of ENUM

Panellists (in alphabetical order): . Denesh Bhabuta 
. Patrik Fältström (Cisco)
. Ondřej Filip (CZ NIC)
. Lino Santos (FCCN)
. Robert Schischka (nic.at GmbH)
. Peter Szegedi (TERENA)

The panel introduced themselves.

Niall O'Reilly asked for questions.

There were no questions.

Niall asked if Denesh Bhabuta had a question for Lino Santos.

Denesh said it was a general question for the panel. He said he's recognised some of the names mentioned when he was in Portugal and the only one that he knows of that does VOIP is Soncarbo(?). He said he sees similar mistakes where the trial goes through, the telcos sign up and then nothing happens. He asked the panel what their plans were to get ENUM going, not on the private interconnect side, but from the user ENUM side.

Lino replied that there is no plan really. They are trying at FCCN is to see if they can benefit the users by aggregating the functions of the ENUM service model as much as they can. Some operators say they're interested but the best way to implement this is to delegate all their numbers and they'll do the rest. FCCN is trying to disaggregate as much as they can. He added that there's a market for registrars and ISPs to provide tier2 nameservers. This must not be the usual players. It can be everyone. The user would like to chose which registrar and which nameservers they want.

Peter Szegedi said they were thinking along the same lines. If you pay attention to validations, registration nameserver providers and the number holder can be all different roles and there may be an open market. Reality has shown that this model doesn't fly: The number, the servers and customer all belong to a single entity. Then the validation is simple. He said that's the only thing that they've seen to work. Reality has shown that the service providers hate the idea that the customer calls them that their line doesn't work and now they have to figure out what has gone wrong: DNS, validation, nameserver, etc. That's part of the problem: It is a very complex product. He added that their lesson learned is that nobody buys ENUM. The best you can get is that someone asks you for a phone service that has some characteristics, like ENUM.

Ondrej Filip said he could confirm this. They had a very similar approach and ended up with the conclusion that there are three types of customers: there are a few geeks, 10-20 people or so. Then there are universities, they do the largest portion of registrations, about half a million. And the biggest part were validated and covered by VOIP operators. He said he didn't see a demand for splitting those roles.

Peter Szegedi said he fully agreed that a single point of ENUM management is good. That's why it works. He added that it was a ‘chicken and egg' problem. You want to convince people to take up ENUM but have nothing to show that works. If you start with an ENUM trial service, you have something to play with and you can hopefully convince everyone in your country that you want to start with it. An example is Norway, maybe they can cooperate with Sweden on this.

Patrick Faltstrom said they have an issue in Sweden with the use of ENUM, the same as everyone else. Providers of services see a connection between their service and the telephone number that it connects. In Sweden this has stalled because at the moment it's impossible to register things in ENUM and has been like that for few years. ENUM is not only interesting for telephone but also for other things you want to identify a telephone number for.

Denesh said he must be one of the geeks mentioned because his seven-week-old daughter has an ENUM based number. He said he spoke to the Indian government about their ENUM. They want to sort out their number portability first before ENUM. He told them ENUM was a good solution for number portability. In the UK, carriers are looking at ENUM for number portability. In Portugal, there seems to be no number portability. Rather than what the UK did with two databases, a private/carrier and user database, everyone should think about having one database.

Lino said they have a portability service between operators, but it doesn't use ENUM.

Denesh asked Patrick if the idea of ENUM was that there were no boundaries with telephone numbers.

Patrick said there were two ideas. The first one was to enable the use of telephone numbers for any kind of service. The second was to enable the End User to have the phone number refer to any provider of a service, different ones for different services. He commented that maybe they've seen too many degrees of freedom, but the architecture makes it difficult to limit.

Robert Schischka said that the attitude of regulators about things like end-points of landlines is that they must be physical and may not be ported to anything virtual and this isn't helpful. He asked who should be the driver if no one creates the ‘killer app'. He said there are VOIP numbers that take years to be reachable from everywhere. In Austria, it took years and it is still not even close to 100%. It is a nightmare to push new operators and number ranges because the customer experience is horrible. He said that everyone worried about spam and speed at the beginning, but that's not been an issue. The largest customer, Manufacturing International, turned off their ENUM. They were using ENUM to connect their phone system. The reason to turn it off was an enormous amount of brute force attacks on their SIP endpoints because ENUM forces your SIP end-points to be public. Brute force attacks are a really bad issue they need to consider.

Ondrej Filip asked Patrick if it was true that there was an interest in other services besides VOIP in Sweden.

Patrick replied that people who are deaf can't use the telephone and they'd like their number to be linked to other services, like chat, etc. He said that at Cisco, they have seen requests for a better link between telephone and VOIP. For that to work the e164 numbers must be available to be used for these deployments. With SIP, you buy a SIP trunk from a telco and they take care of everything. If the telephone numbers are not available, these services will forever use another identifier.

Niall O'Reilly commented that it seems like they're like the road workers on a super highway looking at a pothole and they don't have the right diggers or shovels to know what to do next. He wondered what part of the mindset they have to work at first to try and untie this knot.

Patrick replied that they're all working on new services with other identifiers, if regulators don't allow them to do it with e164 numbers, they'll use something else.

Erika Hersaeus, Swedish Post & Telecom Agency, commented that she wanted to make some clarifications about Sweden and +46. She said that she believed that PTS put two kinds of agreements to .se, how they could become the registry of user ENUM. They did not accept any of these agreements. What she understood, the clash was that the PTS did not believe the state should finance the registry and that was why it led to what happened afterwards. She said their work is to make a concession and they want to see the interest for who wants to run user ENUM in Sweden in spring 2012.

Carsten Schiefner replied that ENUM isn't just about VOIP or addressing SIP trunks. It uses a 10-key pad to identify end-points and access all sorts of Internet services. He asked why do they think no one put all these pieces together to create these services. ENUM is not a service, it is a basic technology. ISPs did the trick roughly 15 years ago: the customer comes with idea for domain name, and the service provider does everything else, name service, website, email, etc. He suggested transporting that idea into ENUM. He asked why no one came up with innovative services around ENUM.

Denesh asked what it was that they wanted to achieve with ENUM. He said it comes down to money. Service providers do not want to offer a service if they don't see a market. It should be ENUM as a technology underlying the service.

Niall replied that it is about finding identifiers to add interesting services to and unifying services by common identity. If the telephone number is not an effective identifier, they will use a different one. It seems a challenge for regulators who control the number space to engage with the innovation or ignore it.

Robert Schischka said there is more than one reason why ENUM is not very successful. To have each country one by one to have their regulator ask for delegation, set it up, etc, does not help the whole thing. He said that he had some customers wanting to use ENUM to interconnect but it failed because it Italy, the outlet works only once a year and in another country there was no delegation. He said they needed six or seven countries but only two were stable. He asked who should offer service if there is no clear income stream. Someone has to pay for the registry and name service to be stable. He added that the current ENUM model was that he pays for someone else to reduce their phone bill. It's a bad business model.

Ondrej Filip said it was more than just money. They offer ENUM numbers for about .50 per year. So registrars cover the costs, but still there is no demand. They don't like the idea of validation and it's annoying for End Users. Lino Santos said that if he was a big company like Google or Facebook, with the instability that's going on, he wouldn't bet on ENUM. It's a chicken and egg problem.

Denesh asked what the answer was then. In Europe, ENUM operators have been left to market forces. On the other hand, they can go for government intervention, like Malaysia. They are more successful. They were selling ENUM as a technology for a service that enhanced the user experience. He asked why they don't want government involved in Europe. He suggested that they might need training about ENUM so they understand it better. He asked why they wouldn't consider government intervention if it's not working.

Robert Schischka said that he hasn't seen a government opening pockets and spending money on it.

Niall agreed with Robert but said there are at least two cases where governments are interested in innovation. It is a buzzword. Ireland for example and France look for innovation at the European level. He said he didn't see ENUM as more special than any other innovation target.

Robert said they may get some funding for trials, but not for production services. That's where innovation stops. Then it becomes day-to-day and an uninteresting business. Everybody has failed to bring an ENUM business case. That's also part of innovation. If somebody comes up with a product that serves a need and demonstrates business case, you can get some money. Just making calls for free is not the ground breaking idea.

Wolfgang Nagele (RIPE NCC) agreed with Robert. He said that in Austria, mobile providers have massive competition. There is almost no reason to ask for cheaper calls. Since the Arab Spring, European governments are keen to spend on free communications and that sort of thing.

Robert replied that they were making money by intercepting these things.

Denesh said that all European ENUM operators need to work together on ideas and not separately.

Robert said the UK had some promising ideas, but suddenly they went completely silent. He wondered what happened.

Denesh said he could only speak of his own views. He left the UK ENUM Consortium 18 months ago. BT was very interested. When the governance body was formed they didn't join, they just kept an eye on it. They set everything up and there were only a few registrations. Everyone was thinking that ENUM was for voice and not for other services. No one implemented them. People liked the idea of ENUM, no one wanted to spend their money. We talked to NHS for using ENUM for information services. There are lots of things it can be used for. End User ENUM doesn't seem to be working. He said this was the reason he asked for the panel and he'd be pleased if operators came together and moved forward on this.

Lino Santos commended that the driver for ENUM is a focus on applications. He asked how they see separating all the components of a service model and putting them all together within the operator.

Robert said that one good reason to separate the roles is because every customer with a number should be able to get an ENUM delegation. One carrier should not be able to block a number range. That's the reason for the separator model. The other way would be for every carrier to have a validation service, they don't have to take part in the ENUM trial, but all their customers can join

Lino replied that the user doesn't have to subscribe to the ENUM service, they subscribe to an application.

Robert said they have the validation step, that's where the trouble starts. The easiest way is for a service provider to validate his own customers. He asked how you prove that you are really the number holder.You end up with a manual process, which is an extra cost and far more expensive than the delegation cost itself.

Peter Szegedi said he wasn't a middleware expert, but it seems similar to the authentication model for federations. Maybe you can apply the same model.

Robert said it wasn't a technology problem. Everyone wants to be the identity provider but no one likes to accept other identities. It's not about a technical protocol, those exist.

Denesh said that in the UK, they would have the application or service providers do the validation. That was the cheapest.

Niall asked the panel to wrap up the conversation. He said there were parallels between the problems facing people who want to offer applications and other federated models. There needs to be identity provider cooperation. This is not the case in the telephone number world. So there is no effective parallel there. The other parallel is social networking, where the identity is email and validation is sending an email. The email infrastructure is there and other services build on top of it. The choice is: Either someone comes up with something compellingly based on phone numbers or we stop working on this ENUM thing. He asked the panellists what they thought about dismantling this and how long they should wait.

Denesh asked if the ENUM Working Group could contact all ENUM operators and let them know that they need to join up thinking and then wait until the next RIPE meeting.

Niall said that Denesh sounded like he was volunteering for an action item.

Denesh agreed.

Peter said they could help.

Carsten said they have three operators sitting at the table here already.

ACTION ENUM-AP-63.1: Denesh Bhabuta, Peter Szegedi: Contact ENUM operators "brainstorming among ENUM operators".

Carsten Schiefner said they would skip the “short news” agenda item. Niall has sent out his slides. If there is an interest, they can go to the list.

F: Short News

F1: enumdata.org update (Niall O'Reilly, mailing list only)

(SKIPPED)

G: Discussion on Plenary presentation:

[Placeholder for 'G': nothing planned this time]

Nothing discussed.

X: Interaction with other working groups

Nothing discussed.

Y: AOB

Niall said they would move the enumdata.org website from Kim Davies to hosting by the RIPE NCC.

ACTION ENUM-AP-63.2: Niall O'Reilly: Move the enumdata.org website.

Z: Close (5 min)

Summary of action items:

ENUM-AP-63.1: Denesh Bhabuta, Peter Szegedi: Contact enum operators "brainstorming among ENUM operators"
ENUM-AP-63.2: Niall O'Reilly: Move the enumdata.org website.