These are unedited transcripts and may contain errors.
Connect BoF session
15 May 2014
9 a.m.
CHAIR: Good morning. Really great to see so many people here this morning.
It's nine o'clock, I only had two cups of coffee. I don't know how your
levels of caffeine is yet. So, please forgive if I can't make myself
understood, there is a reason for that.
So, this is the first time we're doing a Connect BoF and I just want to go
over the agenda.
So, we have 90 minutes and during those 90 minutes we're trying to go
through this ?? I'm going to try and work you through this programme. Of
course we have some ideas in our head about how long a time it will take but
I think one important thing that we need to talk about today is the Charter,
because the whole purpose of this 90 minutes is to talk about how do we move
this group of people into becoming a new voting group, do we still think
it's a good idea to do that? And then we have in the interconnection topics
we try to make a collection of presentations that we, in the organiser
group, believe will be suitable content for a future Working Group to talk
about.
So ?? and what will happen is that we can cut some of those sort if we run
out of time, because we all have this really great and long discussion of
the Working Group Charter and it's going to take forever to agree on that
one.
So, that is my welcome.
I was just going to go present what it is that we ?? or talk a little bit
about what it is that we started out thinking that we were doing in Athens.
We closed down the EIX after many years of very good work because we wanted
to broaden the scope, and we were a couple of people who sat down and have
been talking frequently over the past six months about what do we want to go
moving on from here and one of the things that were really important for the
was to take the whole discussions about interconnection and move it not away
from the EIXs but realise and acknowledge there that are many other ways and
many other connections taking place in the Internet that do not only take
place on an chairs. Internet exchange are still important and we'll hear
what AS about those and we'll hear the news about the internet exchange, but
we will also have a try and have a look at the other ways that we are
connecting the networks today. That's what we did.
One of the things we did was we tried to come up with a new Charter. I sent
it out on the mailing list. There's been a couple of comments on it but I
really hope that there are some people here who have filled up on the
caffeine and then are able to talk about and discuss and comment and have
feelings about what we are going to talk about in the future, possible
future Working Group about interconnections.
That's a lot of text. Do you want a few minutes to read it?
I tried to divide is into what we would like to do, so I think that should
be the first part of the discussion and the group's suggestions is that we
would like to make sure this is one of the places where we discuss the
interconnection and that can be in any layer in the OC model and in any
other layer that any anyone else could come up with.
We are a bunch of technical people. We are a bunch of people who talk to
other people to make this work and we are also a bunch of people who take
care, who follow what is happening on the Government and the Governance
levels, and all this need to blend in for us to still continue making the
Internet work the way we think it should work. So, that's why we will need
to talk about it and we think that this is a good place to talk about it.
We also want to make sure that there is a big group of people who can tell
other people about how important the interconnections actually are for the
Internet because if you move a little bit away from our community, I feel
that there are a lot of people who believe that you know, in Denmark where I
come from, a small country, the Internet is in Copenhagen so if you need
anything with the Internet in Denmark, you just go to the gateway in
Copenhagen, and the people in this room, we all know that's not true, but
step two levels away from us and everybody believes that's how it is. So we
need to make sure that people who are trying to rule or govern or have
opinions that can effect our community, our industry, that they know how
things work.
And that leads in, that's actually the same thing about the next line.
Those people, we need to tell them how the world is. And finally, third way
of saying the same thing. People should know that if we have questions
about interconnections, the people in this room are the people to ask.
Would I like to open for comments on these things now.
AUDIENCE SPEAKER: Patrik Falstrom, NetNod. Do you think people first of
all in this room do agree on what IP interconnect is? Or is that a term
that needs to be defined before you, as part of the Charter? And the next
question is, given that this room do have the same view, whatever it is, is
it part of this Working Group to try to explain to the world that that is
what IP interconnect is and are we sure we are not trying to boil the ocean
here?
NINA BARGISEN: My personal view on that is I'm not sure we need to define
it.
PATRIK FALSTROM: Let me be very precise. I think many people this this
room, specific ISPs believe the IP interconnect has to do with a changing IP
packets between two organisational entities. While in some other cases, we
definitely see that IP interconnect has to do with explicit change of
services in high layer value chain in separate channels where you have
multiple virtual paths where each virtual path has reeached kind of traffic
much higher up the value chain and basically the IPX model. The question is
which one of two is this group going to talk about? Is it both? Is it to
have a discussion on whether we should have in the world one or the other,
or blah blah blah? I'm a little bit nervous that if we are not resolving
the issue on what this group is going to talk about we will have a very big
elephant in the room that no one will talk about.
NINA BARGISEN: I understand your point. I think we should talk about the
elephant, and if I'm understanding you correctly, that means that we should
not limit ourselves to talk about the Internet IP connections. But that is
definitely open for discussion.
PATRIK FALSTROM: As long as we agree.
NINA BARGISEN: Anybody else on this?
AUDIENCE SPEAKER: Morning, I just had my first cup of coffee so I can now
say something. I have no idea... Who am I? Remco, work for Equinix.
I think definitely we should talk about the elephant and at the same time, I
slightly disagree with Patrik in the sense that I don't feel we should lose
ourselves and lose this morning in trying to define IP interconnect exactly.
I think we should just take the broadest possible scope, whether that's just
exchanging packets or exchanging services, that's a matter of semantics at
some point, because we're covering layer one for eight and I'm pretty
confident that at east from layer 6 up, packets don't really matter any
more.
PATRIK FALSTROM: Just let me clarify, I want it in a Charter discussion
that we don't sit here in the room when we're leaving with different views
of what the Charter is saying, that was the only thing. I was not
questioning whether we should define or not.
AUDIENCE SPEAKER: Gordon Lennox, no affiliation. I disagree with Remco
saying high up it doesn't matter. Services are regulated. So, if you are
doing certain services they are regulated as independent down below. The
European Commission has done work on interconnection, which confused the
hell out of some people from the internet world, because they thought they
were trying to connect networks like what the Internet does. In fact, they
were trying to talk about how traditional telecom services no work over what
they call next generation networks, so there is a significant body of work
out there on IP interconnection, I'll put some pointers on the list for
people to look at. We have to be clear. I talked to Patrik about this this
morning. I am quite comfortable if we talk about both sides. What I'm
uncomfortable about is if we get confusioned, if we have one idea of what's
happening here and people elsewhere have another idea. We have to be clear
what we're talking about. If it's both, great, but let's be clear about it.
NINA BARGISEN: A suggestion could be if we agree among ourselves, if we
call Internet IP connections for IP connections and then there is the rest,
is that a dangerous road to go down or is that a way of clarifying what
we're talking about when we talk?
AUDIENCE SPEAKER: I think that's going to confuse the hell out of a whole
different group of people.
NINA BARGISEN: People in this room will be clear. I think that between
Gordon and Patrik we identified the sort of the two broad ideas of what this
entails and if we can get those into the charter we should be fine.
AUDIENCE SPEAKER: I have got two points. I'll only make my point about
this first and then my boss will probably disagree with me.
I'm quite comfortable actually with this level of definition and I think
there are enough people in this room who keep coming to these, whether we
call it the the connection Working Group or the Connect BoF or not and I
think we will fill this Working Group with relevant stuff. So, I'm actually
not too worried about trying to define the scope of the Working Group to
that point.
Now, then I have additional point about the Charter later, but I'll let my
boss speak first.
AUDIENCE SPEAKER: Kurtis Lindqvist. So, RIPE Working Groups, in most
cases, are forums for discussion on a given topic. Some RIPE Working Groups
create policy. I don't think you want to do policy here, right? There is
no one going to do any interconnection policy proposals that run through the
PDP. So, I don't really understand this discussion. The meaning of
interconnect is what will be on the agenda if there is a presentation of
Interconnects of X 500 over IP, so be it, it doesn't matter, if that's
accepted by the Working Group Chairs as a topic, then that's on the agenda.
I think we're rattling.
AUDIENCE SPEAKER: I think you covered most of what I wanted to say. When
we talk about layer 1 to 8 I think we have to acknowledge that we have
expertise in the room on certain layers and not on other layers which
doesn't mean with he have to limit our scope. But I think the elephant in
the room is that we should just be aware that in certain areas we just know
how it works and we have all the knowledge in the room, if we can continue
sharing this. In other areas, we might have to learn and build up
expertise, which is perfectly fine as long as we're aware of it, I think.
NINA BARGISEN: Paul, I actually see that as one of the purposes of the
Working Group.
AUDIENCE SPEAKER: Mike Hughes. Former Working Group Chair of the
predecessor, but I think that's irrelevant. But I think what I'll do is
just have a quick method discussion here. One of the reasons we chose to
dispand EIX was because we had a very functional set of forums, such as
Euro?IX and the peering forums which kind of supplanted or gave greater
focus to those individual areas and effectively those were very successful
and that was the reason that we chose to dispand the IX and I think what we
have got here is the opportunity to go and deal with the wider scope and
deal with, reach a greater audience to take the things that are of general
importance rather than ?? I think the problem is you can become intro aspect
I have quite easily or very sort of inward looking and particularly when you
are in front of a specialist audience, and by coming here, what we have
actually got is the opportunity to go and take the things that are happening
from the more intro aspective groups and put them in front of a wider
audience, discuss them in a wider context and I think as Kurtis was saying,
allow ourselves to be more general rather than specific. So, yeah, I'm just
trying to say don't tie one hand behind our back with this, I think. Barge
thank you. I'm not sure if you wanted anything changed on this first list.
AUDIENCE SPEAKER:
NURANI NIMPUNO: I like the Charter. There was one point that I would
really like to see in here that I don't see and I think one of the things
that did work with the IX Working Group was that it also was, and I'm not
sure how to put it, but it was a forum for people in the interconnection
space, if you can say that. I think you'll get that anyway, but I think
that would be neat to have as part of the Charter.
NINA BARGISEN: Yeah, I agree, it was so in my head that, our in our heads
that that was of course what we were talking about, so we didn't write it
down.
NURANI NIMPUNO: It sort of assumed right, but it's I think it's nice to
specify it as well.
NINA BARGISEN: Is anybody taking notes about this by the way.
STENOGRAPHER: Yeah, me...
AUDIENCE SPEAKER: Mike Hughes again, Nurani got what I was trying to say
in one sentence: Have somewhere that anybody can come and that's what it's
about. And it's that forum, because of course, the peering forum is for
peering coordinators and Euro?IX is for people that run classic exchange
points for instance.
NINA BARGISEN: Okay. Good, I'm going to close the, what we want to do part
and say I think we have an idea now about how to amend this into what the
comments have reflected.
So, the next thing we talked about or the next thing we can talk about it is
how are we are going to do it? And it's quite ?? there isn't really obvious
things that everybody is assuming this is what we do at Working Groups which
is we present and discuss at the meetings, and then we have discussions on
the mailing list, but I think the two next points is probably something that
we need to talk about it that we will be discussing even more later. Which
is we have to make sure, or we need to take care of what is going on in
other Working Groups relateed to this, because this Working Group turns out,
when you look at it when we broaden the scope as we do, we actually have
overlaps with cooperation, specially, but also perhaps other routing, I
don't know, whether we put in databases sometimes that effects also how we
interconnects because it affects how to fill the between the networks and
there is a lot of things that are actually also discussed in other parts of
the community.
So I think it's very important that ?? and that's why we put it in as an
activity, that there are some people who, you know, try to keep the overview
of oh, that's actually relevant for us and do others facilitate that
different Working Groups talk together or whatever or how we can fill out to
make that work.
Another part is again is actually a lot of it is relevant for what's going
on in the governance space and the regulation space. But it's also that you
know, we might try to make sure that we are also heard, that the experts on
how the networks are connected to each other are heard when somebody is
discussing this exact point somewhere else in the world.
So that's the thing behind the last bullet on the screen.
So, comment on this part... silence is agreement, you know that? Awesome.
... oh, Paul.
AUDIENCE SPEAKER: Paul Mandl from Google. I'm trying to phrase this.
So... I would like to reflect the last bullet point on the left side,
actively acknowledge pace support interconnect related questions.
NINA BARGISEN: Oh, but we always cross that.
PAUL MANDL: Yes, but I don't see it much reflected on the activities side.
So just to make it very concrete, right. What I would like to see when I
come to a RIPE Meeting is of course and update for Euro IX, I'm also looking
for of course. But, for example, I would also like to see something from
Peering DB dotcom, I don't see anybody in the room from peering DB, so we
need the practical stuff. We have all this political stuff, which is fine
and we should of course discuss and cover, but then just for practical work
purposes, I think that should also be defined as in scope, so we don't
forget about it.
NINA BARGISEN: Yes... I think you're right. You are taking notes Remco?
We'll try to get that into your ?? worked into the activities as well? But
we probably just need to mention, like, tools and whatever ?? we actually do
have a presentation about that later, but...
So, yes, I think we also came up with just a brainstorm of the topics, and I
think it's very, very important that we see that we can talk about all of
this, but we don't want to limit ourselves so this. Again, we have the
broader scope that is interconnecting networks. I don't know there much
discussion about this, but please say stuff, if I'm not hearing anything
it's because you love it.
Martin Levey, cloud flair. I want to get a couple more people at the mike
here only because this is a bullet item list. We saw it six months ago in
Athens, approximately. But we never put any order on this, and so, what's
important and why I'm at the mike here is to say is there an order where
this is absolutely important, less important, focus on, no focus on? The
one thing we have heard is we want to hear about Euro IX updates, the
question about whether there is anybody from peering B D inside the room,
we'll wait on that one, you'll get a slide on that, I know that. However,
what about the other things? Where are the under sea cables? Are they
really at the bottom of this list, is it really ordered from top the bottom
the way we want to see it. To be absolutely honest, the things that come
into this group that change and what brought part of the conversation is
bringing things like regulation and policy into the discussion where it
didn't have a place inside the IX group. But for me, I take that choice and
data thing and I'd drop is a few notches, but that's just my personal
opinion. I wonder if anybody else has got an opinion adds to what they want
to do with list and if there is an ordering as to how the time is used. I
am trying to rightly the audience and there is no one standing behind me.
NINA BARGISEN: There is, Rudiger.
AUDIENCE SPEAKER: For the record I sincerely apologise for that. Rudiger.
RUDIGER VOLK: Okay. I'm kind of confused about some of the lists. On the
other slide saying layer 1 to 8, well, okay, interconnect for the Internet
on layer 3 is handled in routing. Disagreement? Oh...
NINA BARGISEN: Rudiger, that's why I was talking about the overlaps to
other Working Groups.
RUDIGER VOLK: So what are you talking in ?? what is this group then supposed
to talk in on layer 3? Talking for the Internet. For this list, I'm kind
of confused. Are you going to compete with the mobile world discussing what
they are doing in voice interconnect? Good luck...
NINA BARGISEN: I think, let me try to answer you in what, at least I was
thinking putting this list together. I think it's important for people in
this room that work with interconnecting networks to know to know what the
discussions that is going on in the mobile world about interconnecting their
IP networks. Because, at some point, what we see is that we're using the
same protocol, we are being subjected to some of the same regulations or
regulation ideas for this as they are in the mobile world, so it means
something to what we do and to what we decide to do in our day job about
interconnections. It means something that we know what is going on outside
our own little scope, our own little world. And that's why I think that
this forum could be a really good place to make sure that we have a
possibility to educate people working with interconnect connection on the
Internet to know and understand how the interconnecting in the mobile world
because mobile world tomorrow also want to say high we're also the Internet,
because we're using IP. Same thing with voice, with video, with special
services, with all the things that we fear when we read about them in the
regulators papers because the ideas that are coming out of there is coming
out of the carry yen, the telco space and we as true Internet people shudder
when we hear those words. And I think it's very important that we know what
is going on on in those spaces because they are trying to influence the same
people that we're trying to influence in order to make this Internet
continue as a free and open Internet, which is one aspect of what we do in
this room. And now I'll open ?? stop talking and let other people talk.
AUDIENCE SPEAKER: I'll shoot first, Mike Hughes again. Voice is an
application. That's T it's not special. It's an application. That's my
opinion about voice. But, therefore, do we want to discuss all sorts of
other applications in well you have hinted you may do actually so you have
stolen my thunder there, but do we want to do that? And to what extent do
we want to do that? So again it's probably Patrik in saying there is an
elephant, how big is the elephant? Do we care and do we want to discuss how
applications are impacting interconnect or do we just say it's all just
bits?
AUDIENCE SPEAKER: Kurtis Lindqvist. I went out to get a coffee and I
walked back in and for a moment I thought I was in the an i.e., it meeting.
RIPE Working Groups are not prescriptive. Again we're talking about what
could be possible topics of interest. It's not that we have to create a new
study group to study the IPX standard for voice interconnect. It's all
applications. Voice data ?? and to answer Rudiger's question about routing,
routing is a component of interconnect ?? how to create interconnect traffic
and applications. Do we want to compete with the mobile world congress?
Heck, yes, if they come here they should be allowed to. It doesn't mean
that anyone has to come here, if anyone has an interesting presentation or
topic to discuss, we can discuss it. So be it.
NINA BARGISEN: Thank you.
AUDIENCE SPEAKER: Victor, Dyn. Although my current capacity this is not
necessarily what I look at. I think there is a lot of value including the
IPX GRX items. Maybe we don't have to call it voice specific. I agree it's
an application now. Given the time I spent in this space previously I can
tell you I think there is a lot of value for having a lot of folks who deal
with those interconnects to have a wider exposure to the audience, that
would probably ex you'd the opinion that there is not a lot of skill sets
there in terms of how to do that well. We pretty much have a maturing set
of folks who typically did voice. And then they have kind of grown into
that. So I think there is a lot of value to have those discussions here. I
think there is a lot of wisdom within the other groups that might be exposed
in this facility here to be able to talk about those. There is a lot of ??
anyway, that's a bit of an opinion. In terms of the ordering that Martin
talked about, I don't really care per se, but you might want to put that
down the list so that people don't get shocked and think that that's one of
the two most important things we're going to talk about. The position is
maybe don't talk about voice and data maybe is not important. It's how the
Internet works is maybe more important.
PATRIK FALSTROM: I think Kurtis brought up a very good point earlier. As
this group is not going to talk about policy, I think we should give the
freedom to the Chair and to the Working Group to agree on what kind of
interest interesting topics there are to discuss. Given what I hear the
consensus is leaning towards ?? although, I think it's your job to call
consensus here ?? is that this Charter is not supposed to be used as a tool
to say no to someone that brings up something that might be interesting for
whatever reasons.
NINA BARGISEN: Yes. Nurani.
NURANI NIMPUNO: I think if the sub?title said, in order of importance, this
Working Group will only discuss the following topics, yeah, I'd have a
problem. But that's not what it says. So let's move on.
NINA BARGISEN: Thank you. And if it was prioritised I would have used the
number button when I did the list.
AUDIENCE SPEAKER: Michele here, from my aspect I think all of these issues
are real issues in Africa that we're actually facing, so I kind of agree
with Martin, especially with the bottom bullet, I personally feel it should
be at the top. But one of the things that I'd like to see is how can we
probably extend this kind of Working Group back into Africa, because we
definitely don't have these discussions that happen at the moment and
currently it's more being driven by Government regulation, specifically
actually the voice one funnily enough.
NINA BARGISEN: That's interesting, so what you are actually saying is
perhaps look for opportunities to take some of what is going on here and
take it with you back to Africa?
AUDIENCE SPEAKER: I'm trying to think of ideas on how we can actually
build forums like this in Africa, and maybe it's a learning that we can take
from you guys.
NINA BARGISEN: Greathank you. Right. I'm not so well into the whole
formal thing, but Patrik just told me to call consent, so, yes, we are
moving on, we have a Charter with a few amendments that will come out on the
mailing list, right? Yes.
RANDY BUSH: Can I just say one word, Michele, AFNOG, come to the meeting in
Djibouti, week after next.
NINA BARGISEN: I think that was my presentation. What I put in now for the
next thing in the agenda. Let me go back to that one, is that we have under
the Charter we have been discussing the regulation a little bit, and so we
have a sub agenda point in the Charter discussion, which I forgot to put on
the agenda, but Gordon has a small presentation on BEREC, but we can take
that now if that's good or we can do it later. We'll take it now. Good.
SPEAKER: Hi, I have been wandering around these rooms and corridors with
this in my hand and frightening people with it and it the end, Remco asked
me to talk about it publicly.
BEREC and RIPE.
This is either a test case or a heads up or maybe both for this new Working
Group. And the Cooperation Working Group. BEREC. BEREC is the body of
European regulators for electronic communications, established 2009, got
really operational 2011. So it's a fairly new body. It follows on from an
older body.
It's ?? this is some text from the website. It talks about what they're
doing. BEREC is important, because it tells the other regulators how to do
their job. The legislation says what they have to do, what BEREC says is
how you do it to try and apply the regulation consistently, so BEREC is an
interesting organisation I think.
One of the things on the website, national regulatory authorities and the
commission have to take utmost account of any opinion, recommendation,
guidelines, advice, whatever, that comes out of BEREC. BEREC is
interesting. BEREC have been working on network neutrality since basically
it started. They had a recent public consultation. Monitoring quality of
Internet access services in the context of network neutrality. The link is
there, it's still on the home page, it's fairly recent.
By the way, network neutrality is also being discussed in the European
apartment recently. It's based on a proposed new regulation from the
European Commission. We have had the vote on first reading, we're now
waiting for the Council to come back with their opinion on the proposed
regulation, and then I would expect, I may be wrong, but I'd expect we'd
have the regulation adopted before the end of the year. Meanwhile, as I
say, BEREC has been working quite heavily on what network neutrality means
in terms of the regulations and how they apply it.
The EU approach network neutrality, two basic pillars, competition
transparency. The idea is if consumers have the right to choose and they
can only choose the appropriate information. So, the idea is there'd be
competition and you can choose one access provider and another, based on
what they are providing.
Following on from that, BEREC has developed some terminology, some concepts
and that's what they are working on. They have had this notion of the
Internet as one service and specialised services is all the rest. They are
talking about quality of service metrics, and access, these are kind of ITU
quality service met tricks, they are talking with monitoring tools,
services, they refer in the report there to what the RIPE NCC is doing with
Atlas and they are talking about everything from access services through to
IXPs and IPXs and so on. It covers a whole scope.
So, BEREC's work affects the RIPE community, individuals, organisations.
One important point to remember is BEREC, like the commission, has its own
timetable, it doesn't work to our timetable. Sometimes we have to work with
their timetable, if we have to some effect. Too often, I feel, we're
playing catchup. The big question is given the significant amount of work
going on out there, defining new concepts, new terminology, how they are
going to regulate the Internet in the future, which Working Group should
deal with this, particularly looking forward to London, which in some ways
might even be a bit too late, but at least I'd like some clarification on
how to move forward on this. So it's just questions. Thank you. That's
nice, nobody at the mike, I'll go away.
NINA BARGISEN: But, people we need to answer his questions. Personally, I
would like to see people in ?? I would like to see this discussed and at
least with people from this Working Group, but I really would like input
from other people, if you feel that we can, as a community in this room, as
a people working with interconnection, can we face that attacks of following
the BEREC, or do we have to ask Maria to do it on our behalf and give heroin
put, Maria and Meredith and a lane, or what are we going to do here?
Kurtis?
KURTIS LINDQVIST: BEREC has actually been quite open in the past and
organised workshops. Most people here have been at least, their companies
have been represented in BEREC meetings. They have been actually quite
responsive, for being in the EC they have been been very communicative and
open what they are doing and what they think and they are taking input from
a lot of people.
NINA BARGISEN: Invite them. That's a brilliant idea. Yes, that's what
we're going to do. Yes. Excellent. And then we take it from there. And
we have to remember that we are not only supposed to be talking about this
at these meetings, in this meeting, on a Thursday morning at a RIPE Meeting,
we are also supposed to be talking about all of this on the mailing list
everyday.
AUDIENCE SPEAKER: Alan, no affiliation. Invite them, right?
NINA BARGISEN: Make them come is the next task.
AUDIENCE SPEAKER: Maybe you should also considering to them. The story
about the the mountain and Mo success. One has to approach the other and if
they don't come, we'll have to come to them, that's my proposal.
NINA BARGISEN: Yes, but I think we can agree that is what is going ?? we
need that to happen. But, I think the question is also: Is it us here who
try to go to them if they don't want to come to us? Or is it the next
Working Group starting at 11, or is it a joint venture?
AUDIENCE SPEAKER: As long as it happens,
NINA BARGISEN: Who cares, I agree. Good. Silence is consent, so ?? okay,
Remco.
REMCO: I think this is an excellent example of how this Working Group?to?be
is going to cooperate with other Working Groups like the aptly named
Cooperation Working Group. So, I think this is going to be very much a
joint venture.
NINA BARGISEN: Okay. I think we need to move on. Otherwise you will not
be hearing the last parts of the agenda. Up next is Michele, and now we are
moving onto the examples of what we think the Working Group should look like
in the future.
MICHELE McCANN: Good morning everybody. So what I'd thought I'd do it kind
of highlight what I have mainly discussed this week with everybody from
Africa. So you'll see in the top corner there that's Zuma's house house, he
did spend millions and millions of dollars on it, hence why he looked so
happy. Oscar Pistorius, is he crazy? I'll leave it up to you guys but I
don't know any individuals of yourselves who would start shooting at a
bathroom door. I'll leave it up to you guys on that one.
However, the relevance of this is that for the first time in after cave we
are actually starting to see this kind of content being available online,
which is really exciting, because we probably have one of the largest mobile
users in the world.
So, going on to that is, this is an advert that was recently played, so we
have kind of a huge mobile war interconnection happening at the moment, and
if you have a look, M, it N did a full page advert in our national
newspaper, and cell C responded with this kind of advert back, which I
thought was quite clever, and the response after that that's now happened is
Vodacom is actually starting to lose users because they also haven't had
their full page advert. But what's happened under the interconnection rates
is previously we used to paying about 7 dollar?cents per minute and now it's
actually dropped to 4 dollar?cents, just purely through Government actually
funnily enough having to step in and saying to mobile operators your
interconnection costs are too high.
So, there is our innovation for Africa. We love fibre baskets and we are
very good at it. So there is an example of one, so when you guys come
through to Africa and buy one of those, remember you're on the
inter?connection Working Group.
So, essentially, actually interesting thing around the stats is probably the
education side it have and the majority of those breaks that you see under
the fibre side is actually the guys physically digging up the roads,
thinking it's copper and actually taking out the fibre, bus costing the
operators forth nightly union of revenue as well as it's kind of slowing
down in their words, fibre to the home roll?outs.
So, Africa itself, you guys have seen me show this map quite a couple of
times, but this address is what Martin mentioned earlier in terms of the
cross?border connectivity issues, Africa itself is one of the largest
continents out there which obviously makes it incredibly challenging to roll
out fibre networks throughout the whole of Africa. We have, however,
advanced quite a bit where we now actually have four major hubs, where we
can actually start concentrating activities around those, which then helps
specifically on IXP growth. Data centre Doe employment, etc..
Key drivers. There is loads of business in Africa, and why we kind of hang
quite heavily onto the mobile side of it is because a major product in
Africa is mobile money. Example of that is Impeza which was launched in
Kenya, with he see a the lot of our banking transacstions via mobile and
money lending happens via mobile transactions. As you can see in terms of
the percent ance it's one of the largest growing applications.
Do we have capacity? Yes we do have capacity. So we have loads of fibre
cables, or under sea cables coming into the country. Sea come just
announced this week that they have upgraded to 12 terabytes worth of
capacity. Which is great news. However, the pricing side it have which is
what you have taken out of this meeting this week is a number of you are
getting very interesting pricing. Which is then stopping content, for
example, coming into Africa. So, from a pricing perspective, I will say it
openly, we should be and pretty much depending on how hard you negotiate
with these guys, are at the same levels of Australia right now.
/TR*ES relevant fibre, yes we do have it and there has been extensive
/ roll?outs through companies such as liquid, CM C and sea come who have
/ mainly addressed a number of cross border activities. As an example of
/ this we have seen Malawi has joined and /PREFP were paying $3,000 per
/ 1?meg to be able to access international content. And by them being able
/ to come cross?border, they are now actually paying $300, which is still
/ very high in your terms, but in their terms it's made a huge difference.
/ Malawi itself is getting 40% of its content directly from the South
/ African Internet exchange.
Has it had and economical impact? Absolutely. So, it's new, it's emerging,
and it absolutely has had an massive economical impact and we have seen this
growth happening across the boards.
This is an example of Teracos interconnection. So the one challenge that I
actually found from a data centre operator's point of view is the
understanding of who is in the data centre and who I can physically
interconnect to. We have had to spend many hours to educating how are
market to actually say there are other members within the data centre that
you can actually interconnect and hence while you will see there's been
quite a big spike in pretty much this last year.
So why that is there had to be a lot of evidencecation around this? This is
an example of a service provider data centre that happens in Africa and very
much in South Africa, and typically what was happening is if I was hosting
in one of those service provider data centres, and I wanted to connect to my
potential client next to me, I would have to pay per meg transit fees within
the data centre. So it sounds really crazy, I met with the Apple guys
yesterday and they said are you nuts? I said yes, this does happen.
Previously, I worked in a company called eye BIS and we hosted in that
service provider data centre and wanted to launch a product with my
neighbour right next to me in the data centre, and I actually couldn't,
because I couldn't afford the in?data centre costs. Crazy. So starting to
we are getting some neutral facilities, Teraco being one, in Kenya being
another one, east Africa data centre, where typically you are charging for
the actual cross connect itself, so cable management fee. And this is
caused a lot of politics, would I say, because your service providers
generally used to earn up to 25% of their revenue out of end to daren't
transit. We have a challenge between us and Kenya to actually try and break
these rules and put policies and precedence in place on how you should
interconnect within the data centre.
So, example in terms of savings. So this is a major banking client of ours
who previously used to ?? it's 50 thousand dollars a month for three hundred
MEGS. Purely by peering, he's been able to save that over a very short
period of time which also gives you an idea of the African transit costs
which are incredibly high, as well as the lack of IXPs that are available.
Example, we only actually have 26 IXPs through that entire African region.
So how that is this all happened? This is South Africa and it's been
through deregulation. So a number of African countries are not deregulated.
We had a fantastic company who took on the Government in 2008 and said there
should not be only one monopoly network available and won the ruling. So
now we have, over from a fibre perspective, we have got at least over 30
network operators that are rolling out fibre, connecting on to exchanges and
purchasing under sea capacity.
A long list of issues, I think it's longer than Nina's list. So I bolded
the key ones. But key highlights on this one, and we see it on a daily
basis is the lack of skills, and that's the lack of technical skills in
Africa. To turn up appear sometimes can take us a day, because it doesn't
understand what BGP routing is, he has no idea. It's his first time he has
gotten an ASN number and he is using us as the training platform, which is
great. From an IXP perspective that's our role to be able to do that, but
one of the things I start urging after cans to do is we generally tend to
hide behind forums and complain about our general colleagues and you hear
about it when you are sitting in Poland, rather pick up the phone and phone
that guy and say do you understand what actually BGP means and how can I
help you, rather than routing your traffic all over the world.
So, this is a picture that's been travelling around quite a bit and that's
the mobile wars, so you see that in all the papers everywhere. But,
absolutely openness is really required. We need our Government, once again,
you are talking about policies to take a step back and what they want to
actually do is regulate regional exchanges, and choose who is the regional
exchange, as well as they want to choose who is the regional carrier, which
is crazy. Rather let the market decide who that should be and how it should
be operating. But our Government, which pretty much don't even know what an
Internet access connection is, is trying to create these rules and luckily
ISOC has stepped in on this and is working incredibly hard to try and move
the boundaries on this one.
So, in ending because Nina is going to shout at me just now ?? and I need my
coffee ?? so, in ending, as you have seen, there are a number of
interconnection issues and that comes out and you guys have a lot of
knowledge around this because you have gone through these issues in the
past. One of the key things that us as African need to do we need to stop
hanging on to Legacy methods and revs, let those go and start innovating and
I really like this slide because the 23%, that's MixIt, and that was
actually a South African innovation, which is pretty much, if you look at
Facebook, WhatsApp is really nearly up to the same level as WhatsApp in
Africa and that's being used in Africa.
Questions? Ideas?
AUDIENCE SPEAKER: I have a question, Carsten. I just wonder whether you
could shed some light on your point of view of what's the current state of
the debate in zero rated content in Africa is? Zero rated content, so
content that comes from certain content providers, but not going to be put
against like your kept data rate on your particular ??
MICHELE MccANN: It depends on which country in Africa. Personally I think
that that must happen and orange, funny enough, they had a huge success in
Africa, because what they did was they said, Facebook does not count towards
your data cap, and everyone ran towards that and hence why orange is
actually doing incredibly well in Africa. I have been urging a lot of
mobile operators to try and do that because we all know it's very easy to do
it on radio servers etc., however, they still are hanging on to a lot of
legacy revenues.
NICK HILLIARD: I think this is great. I have been around for a while. 15
years ago we were in this sort of situation in Ireland. We had content
networks being built in the US. The US didn't really realise that there was
anything outside the US and there was a huge lack of native content within
Europe at the time. It's changed dramatically because Europe became more
important to the Internet and I think in time, Africa is going to have a
massive growth rate and you're going to get a whole pile of local content
but that doesn't mean you don't have to fight for it still. You do. It's
going to be several years before you get silly anomalies like you know, the
Kenyan nation newspaper being hosted in Amazon AWS in Ireland, so well done
for coming here, making your case to everybody, and sort of saying that
after cave exists, and it is important to the future of the Internet, to the
global Internet.
MICHELE McCANN: On that point, is one of the things that I need to go back
to, is ?? and I was actually chatting to my colleague Andrew about it is us
as after cans we kind of think we are the only ones who exist in the world.
So as much as I stand here and say we need skills and you probably see us
our ad on the TV we need to ?? we actually start to do some actions and more
of my colleagues need to start attending these RIPE meetings.
(Applause) up next is Bijal.
BIJAL SANGHANI: I'm going to give you hopefully a quick update on what's
been happening with Euro IX and a little bit more.
So, I'm going to skip through some of these slides because you have probably
seen them before and if you want to read them in more detail, you can look
at the slides online. So Euro IX is an association for Internet Exchanges.
We have two forums a year, we maintain a website, a database with tools. We
have mailing lists, we are recently started doing news letters, news letters
are public so if you are interested in getting a copy of the newsletter
which includes news about IXPs and what's going on in the IXP world, then
you can subscribe there. We have Working Groups, taskforces, twinning
programme, benchmarking, I'm not going to focus on that too much.
We have 77 an affiliated IXPs, 53 are those are in Europe, 32 countries.
And 24 from the rest of the world.
We also have a patron scheme and we currently have 11 patrons and they are
there.
And this is the Euro IX region which as you'll see is the same as the RIPE
region.
Number of IXPs in the Euro IX region, the current count is 183. And we have
an interactive map as well so if you want to have a look at that, you can go
down down and see where the IXPs are and link to them from the map.
There has been continued growth in IXPs, as you can see, which is a good
thing. This is just in the Euro IX region.
And the traffic growth, I am going to skip through this, but like I said you
can see these all online, but what we have started to do now is we are
consciously getting in touch with more and more IXPs so that we can try and
get as much data as we can from them. So in the last year, we have gone
from collective monitoring like for 71 IXPs in Europe to 85. So you can see
that's also increased the traffic there.
In Asia, there is 89 IXPs, again, you can have a look at the map to
draw?down. And the traffic rate in Asia. I'm going to skip through a lot
of this because I have got some more interesting things to talk about a bit
later on. Traffic growth in Africa, again, we went from monitoring 5 IXPs
last year to 11, and that's hopefully going to continue to grow, as more and
more IXPs come on the scene there.
IXPs in Latin America. And the traffic growth in Latin America.
IXPs in North America. Traffic in North America.
So we have a number of like I said we have got a number of databases, and
what we do is we have looked at how many ASNs are actually connecting to the
IXPs. I think what's interesting here is you know, 30% of ?? and this is
the Euro IX members only ?? 30% of ?? 3 to 25 ASNs connected, which is the
largest number there.
And who connects to an IXP? This is comparison from some data that we have
taken in 2012 and compared it to last year, and as you can see, there is
pretty much growth in most areas. I think the gambling company has been a
bigger growth there. But I'll leave that to, so you can look at that later.
We have a switch and a router database, and this graph shows the number of
?? well what switches and routers are being used at the various different
IXPs. And also some examples of which IXPs are actually using those.
Route server database. These are the route servers that are being used
within the IXP.
Which operating systems are used.
Other Internet Exchange points. As, you know, I'm giving you Euro IX
update. There is also API X, Af?IX and lack I F.
In 2012 we signed an MoU to form an Internet Exchange federation and the
idea is to have a global IXP database. I'm going to talk about that in a
few minutes as well.
Also, to set some standards for IXPs and best common practices that can auto
bandwidth you had globally, automate data collection from IXPs and plans to
collaborate with external databases.
And this is the IXP board. It's two representatives from each of the
Internet Exchange point associations come together to form the federation
board.
The exciting stuff:
Like I said, we have talked about having a global database and it's
something that one of the things that the federation has really, really,
really wanted to focus on because there isn't a place for IXPs to kind of
own their own data. So, you know, there is lots of IXP databases out there.
So, what we have decided to do is currently develop the current Euro IX
database, and what this is is the only source that IXPs own their own data.
The IXPs data is maintained by the IXPs which means that the data accuracy
is much better, because it's them that's putting that data in directly. I
guess it should have come at the top. It's a single place for IXP data.
It's also great for research and also useful for IXPs members, to you guys,
to seek out IXP data and content. Others will be able to mirror the IX F
data. This is the grand plan. We have the IX F database, with all the ??
and the IXPs feed directly into that. We will be working with peering DBs
using the IXP info into the peering DB and using data centre and network
info from peering DB and use that as APIs.
So, last week in Sheffield we had a Hackathon. The plan is to build a new
database for IXPs. Again, the IXPs will own their own data, and we have
already started doing a massive cleanup between the IXPs that are in peering
DB and also in the current Euro IX database, and we wanted something that
was going to be open and plugable.
So, the IX F DB server is written using Python and YANG owe which can get,
create, update, delete IXP and organisation objects. All interaction is
done? JSON. All nonsense tif information will be publicly available.
Members of the IXP As and others as appropriate will be able to create,
update and delete IXP data from the databaseses themselves.
We have a proof of concept client to interact with this database in Python,
including unit test code. The PHP version of this with unit tests is also
available on GIT hub. This is a work in progress. This was just last week
and I really wanted to give some kind of update to everyone today. We have
migration scripts at the moment to keep the federation, the IXF data in sync
with currently with the Euro IX data, and the code is available on here.
And that's it. Lastly, just the Internet revealed the movie, we have now
got it in lots of different languages and if you are interested in
translating it into your own language, then please get in touch.
Are there any questions?
NURANI NIMPUNO: Just quickly, it's not a question. It's a comment. I
think this project is fantastic. I think there is grade need for it. I am
sad enough to get excited about it myself, but I think there is ?? well, you
know this, but there is a great need for authoritative data on this, not
just in this world but sort of the world out there as well, so
congratulations and good luck.
(Applause)
AUDIENCE SPEAKER: Blake. Just to comment on the geographic distribution
of IX association, there was kind of a gaping hole on North America. You
might want to talk to the guys at hope IX, Martin Hannigan & Co. I know
relative to the other organisations, they are getting off the ground, they
don't have databases and thing yet, but...
CHAIR: The federation also recognise sizes there is a gap and we would look
to see an association in North America to join, so, I'm hopefully something
will happen soon.
(Applause)
WILL HARGRAVE: Hold on tight guys. Whose been to the data centre this year
and plugged a fibre in? Pretty good. Christian you haven't been to a data
centre, come on.
Anyway. Take a look at these SS Ps, they look weird. Is there a little
hole missing? Yes, these are single fibre optics. We launched a new
product. We are delivering 1 or 10 gig service using single fibre optics.
These are ?? they have come basically out of the PON industry, where fibre
count is important and we have started using them for interconnection. This
grew out of a need where we had a situation where a lot of people were
upgrading from 1 gig to 10 gig or 10 gig to 20 gig. Getting into second
cross?connectors is a pain for them.
So, we started offering this as a service for no additional cost. The
advantage that we saved on ODF space, they save cross connect fees. You
never have a problem where you have to swap TX and RX.
And we started doing this at the start of the year.
What are the down sides? It's a weird optic type. You need a different
optic at each end of the link. It's confusing and weird and more expensive.
We expect this will be mainly used by the more included up and more NERDey
operators.
We have some interface descriptions. These are the things you need to order
to connect using this. The cost is obviously not that much. And also,
we're doing a 10 gig, the 10 gig standard, the availability is not as high
as a 1 gig, the things are very easy to get at 1 gig but a 10 gig is rather
less summon.
And that's it.
AUDIENCE SPEAKER: Mick Hughes. A slightly snide comment here. Have you
had a consideration about what this will to do the remote hands, smart hands
revenue for the co?lo operators for swapping TX and RX? You are just
chewing into people's revenues streams here.
That is a joke, I am not sheerious, just for the record.
NICK HILLIARD: Nick Hilliard. This has a really interesting failure mode,
because if you have a cross?connect and you report a flaw or a problem, a
connectivity problem to the data centre operator, and if the data centre
operator is kind of clued up, they kind of poke around with cross?connects
in the meet?me room, and they figure out which is TX and which is RX because
they think in terms of TX and RX. So they put their light metre on and then
they match up, they say okay there is nothing coming in one side and the
other, and then they'll match up the cables the wrong way and your
cross?connect will go down. That has actually happened to us it's a pain.
It's actually happened to us in metro fibre which is even worse.
SPEAKER: It does have the advantage if you measure the light level at one
end you know the I go threat on both ends. Don't need to know two numbers
to whether whether you are cross?connect is in spec. You only need to know
one number and you know the service is working. That's the flip side of
that. But I can see Nina jumping up and down and either she needs the loo
or she wants me to get off the stage.
FERGUS McKAY: Thank you, Will, for the commercial advert for us.
Fergus McKay, a quick update on open IX.
I am going to assume most people know where open IX was created. Basically
it was to try and improve both the IXP and the data centre scene in North
America taking as the general model, Europe.
We got 138 members, people like me are members, people in the audience can
be members. It's a membership organisation of individuals with sponsorship
from companies but they are not actually members. We got 16 certified data
centres in the US and two IXPs, AMS?IX and links. Another five data centre
going through the approval process and two IXPs. The two kinds of
certification we have are OIX 1 which is for IXPs and for data centres I say
OIX 2. We do actually have a database, it's a directory. It's a bit
hard?driven, it does have some broken links on it.
The net effect so far has been that we have seen changes in pricing which
are positive generally, pricing going down but also know that pricing has to
be published openly, there is a lot more transparency in the market which
has helped people ensure that they are all working to an even keel. On the
IXP side, it's tightening up the processes, making it clear what you need to
offer as an IXP rather than just being an ad hoc organisation.
The people who are driving the processes internally, we have got Hank, who I
think is in Chicago today, but Gabe and Christian who don't normally come to
RIPE meetings, but the three of them are all driving the individual
committees which drive the certification.
And if you want to join, it's only 50 bucks, you get a T?shirt, you get to
vote, you get to stand up here and talk, you get to serve on the committees,
you get to have an opinion. Be here on the membership list and there are
private events which are held around usually NANOG meetings.
You want to sign up? There you have. Any questions? No, great.
(Applause)
NINA BARGISEN: So, I think I have some sad news. Because, we are running
really late of time and we do have one more really important matter that we
need to talk about which is what are we going to do with this BoF, are we
going to move forward to become a Working Group? Who are we going to select
for chairs? So sadly, I think we are going to postpone this one to the next
BoF Working Group, because we only have ten minutes left of our slot ?? so,
we had a fight about it just two minutes ago, and you know, she is really
upset. But she bows for the better of the community. So... and shell get a
beer for it.
We are going to move onto the more admin, that's a really hard word to say
at this time. Section do it. Two things in there: Chairs, moving on, and
then do we have consents about asking the Plenary to turn this into a
Working Group or do we continue as a BoF a little bit longer?
So, first, I do believe we have two people who would like to co?chair this
in the future. Do you want me to present you or do you want to present
yourselves? Remco? And Florence. So you come up here and talk about what
you want to do.
SPEAKER: My name is Remco. My day job is not heckling in a microphone even
though some people like to think I do. My day job is also not on the Board
of RIPE NCC, even though that's a bit of a hobby. My actual day job, and
this might surprise some people is I run interconnection for Equinix in
Europe and the Middle East, which kind of puts this Working Group, or this
BoF, right at the heart of the stuff I do on on a day?to?day basis. So I
would very much like the opportunity to Chair this Working Group. That's
really all I'm going to say about it. Thank you.
SPEAKER: Morning everyone, my name is Florence Lavroff, working for Google,
I am responsible for the management of all the private interconnects we have
with our peers in France, Italy, the Nordics, northern Africa and also
eastern Europe. I'm also an ?? where I basically do the same kind of job
and with this, more than five years' experience, I would like to help Remco
here in this BoF. That would be a really great experience for me.
NINA BARGISEN: So, the exciting question now is, do we have anybody else
who wants to be part of this? Can we agree to select these two people as
chairs?
(Applause)
Great. So the last point is, are we ready to ask if we can become a Working
Group? Personally, I think we are. We have some excellent people for
chairs. I think we had a really good discussion about the Charter where we
have consensus about what we would like to do. So, I think what I'm going
to ask is is anybody going to talk against that? Mike?
AUDIENCE SPEAKER: Oh, I'm not going to throw a cat amongst the pigeons. I
think we're not ready. I think there needs to be a wider ?? the Charter
discussion went on for quite a while, and did we get consensus? I'm not
sure. I think we might want to go around the boy with another BoF just to
get a better feel for the types of discussion that we're going to have here.
We have had a taster of what this could look like, we just, by consensus,
selected some new chairs, and I think what would be good is to allow them
the chance to shape this even further before it becomes a Working Group with
that Charter. I don't know what they think about that, but that's my
feeling.
KURTIS LINDQVIST: I disagree with Mike. I think there was a lot of
discussion, but I don't think anyone was actually against the list. I think
it was unclear what the Charter meant but I didn't hear anyone say no. I
think the best way of doing this is to ask if there is consensus for a
Charter and ask if there is consensus with running a Working Group. If you
get yes for both questions you're done.
AUDIENCE SPEAKER: I want to say I kind of agree with Kurtis. I think, in
a way I agree with Mike that you know you should let them start working
things and shaping things, but I'd like to do it as a Working Group. I'd
like them to Chair a Working Group and start shaping this matter, so I would
support turning this into a Working Group.
AUDIENCE SPEAKER: I agree with the previous two speakers ?? Martin Levey
?? and take back what I said the previous time at the microphone. After
being explained that the ordering of the items was not the numbered list, I
take that back, and I'm quite happy with everything I have heard and I said
I agree, I think we're ready to move forward.
AUDIENCE SPEAKER: Paul from Google. One practical question which I don't
know the answer to.
So, if we were to agree on this Charter and put it forward to the Plenary,
is there any way we can clean it up later and phrase it better or are we
stuck with it? Anyone knows?
NINA BARGISEN: Could somebody who knows that answer that question?
SPEAKER: We can recharter, it just goes through a process. We do it
internally.
AUDIENCE SPEAKER: Just to add?on that. I think it could be improved, just
to be clear, so I don't think we have a principal disagreement of on what
has been put down in the Charter, but I think it could be worked out a bit
better way, let's put it that way.
NIGEL TITLEY: I think the group should definitely go forward. Apart from
anything else you already provided some wonderful secret Working Group
material.
AUDIENCE SPEAKER: Remco speaking for the first time as Working Group Chair
elect, I guess. It's BoF something...
Yes, mike, I see your point that we, and to Paul's point, the Charter could
use some additional work. I think we do have ?? well, in other groups it's
called rough consensus, what it means. There could definitely be sort of
refining of what it looks like, but I don't think that should hold us up in
forming a Working Group.
NICK HILLIARD: Success is normally achieved incrementally, not in fell
swoop...
NINA BARGISEN: Right. So, we are going to do the two motions that Kurtis
suggested? I forgot what they were.
KURTIS LINDQVIST: Do you have consensus for the Charter and do you have
consensus for forming a Working Group?
NINA BARGISEN: Okay, do we have consensus for the Charter?
(Applause)
I see most people clapping, I think yes.
Is anybody against?
Do we have consensus to ask for forming a Working Group?
(Applause)
Do we have anyone against?
Great. We are going to do that in the Closing Plenary, and now I think it's
time for coffee.
(Coffee break)
LIVE CAPTIONING BY MARY McKEON RMR, CRR, CBC
DOYLE COURT REPORTERS LTD, DUBLIN, IRELAND.
WWW.DCR.IE
