NANOG 24: NAP BoF

Mailman List Mirror (Read Only)
1

Hi all,

As you may have seen from the NANOG 24 agenda, I'm running two NAP
sessions at the coming meeting.

In the General Session, there is a NAP update, from a handful of NAPs.
We've had to keep this session small, and tightly timed, to fit well into
the agenda.

On the Monday evening, we'll have a NAP BoF session, which will I'm
planning on being fairly free-form, but will consist of:

* an "open-mic" for operators who didn't participate in the general
session to introduce themselves and their IX/NAP.

* a chance to carry out further discussion of any "hot topics" from the
general session - I've even asked for a whiteboard/flipchart.

* a place where participants get the opportunity to meet and chat with the
operators of the NAPs where they participate now, or may be participating
in the future.

* We'll probably head toward the bar at some point too... :slight_smile:

Why have we chosen to split it up like this? Well, in Europe, at the RIPE
meeting, there is an open European IXP working-group for which we get 3
hours of agenda time. Not all the IXPs show up, and still we are
hard-pressed to fit things in the schedule!

So, if there are any NAP/IX operators who would be interested in coming
along to the evening BoF, drop me a mail, or you're free to just turn up
on the night. Of course, we would also like plenty of NAP participants to
show up too! It will also be a good chance to meet with potential new
peers as well...

See you in Miami!
Mike

2

i believe there are/were only four NAPs, so scheduling should not be a
problem.

randy

3

Not!

www.ep.net

Randy Bush wrote:

4

Randy Bush wrote:
    >> i believe there are/were only four NAPs, so scheduling should not be a
    >> problem.

      Richard Irving wrote:
    > Not!
    > www.ep.net
          ^^

Um, does that spell "NAP" to you?

NAP means something specific. I think you're confusing it with "exchange
point". And Randy isn't. Plus, you top-posted.

                                -Bill

5

i believe there are/were only four NAPs, so scheduling should not be a
problem.

it seems that many folk do not know that NSF let contracts involving four
NAPs, Pennsauken, AADS, PacBell, and MAE-East (i hope bit-rot has not set in
so badly i blew that list). those are/were the only NAPs. there are many
more excnagne points.

the essence of what NSF said was that, to get a transition contract, a
provider had to be at the NAPs so others would know where/how to reach
them. what they did not say was that the providers had to peer openly
and provision sufficient bandwidth; but that's another story.

i am sure this is all written down somewhere, which is good as i suffer
from increasing bit-rot.

i think george santayana had something to say about all this.

randy

6

Randy Bush wrote:
    > it seems that many folk do not know that NSF let contracts involving four
    > NAPs, Pennsauken, AADS, PacBell, and MAE-East

As part of the NII, the National Information Infrastructure. The same
document defined NSP, ISP, and IAP.

    > There are many more exchange points.

About three hundred, by the count that Bill, Antony, and I have been
keeping:

http://www.pch.net/documents/data/exchange-points/ep-in-addrs.txt
http://www.pch.net/documents/data/exchange-points/ep-in-addrs.xls

Additions and corrections solicited.

                                -Bill

7

If you recall... NAP's were tiered... 1,2 etc...

So, you now claim only Tier 1's are actual NAP's ?

(not wanting to send reporters to the
phones with yet another mythconception)

  Why don't we take a look at the listings for North America,
from this site....

and we can see where I developed this "incredible misnomer"...

Perhaps we will need people to rename their "exchanges", eh ?

Get Bill M's call on it..... Tiering was his Idea.

8

If you recall... NAP's were tiered... 1,2 etc...

no, i do not recall that. perhaps it is hard enough to learn from
history without rewriting it.

randy

9

Something bothers me about this thread, and I think it is the
assertion that there is a proper definition for 'NAP' as
differentiated from 'IXP' or what not.

It seems as silly as trying to define Tier 1 v. Tier 2.

The NSF's documents do not define terminology or protocols in a
manner like the IETF for the operation of the greater Internet.

I believe the referenced documents discuss requirements for the
transition of the backbone over time, but i don't have them in
front of me, nor am I eidetic like some folks seem.

People should be free to define these terms as they see fit,
with no central authority defining what is what (and I don't
think there should be).

A scientific community will communicate (only|best) through a
common vernacular, but to disallow ambiguity in certain terms is
to require unanimity, which will not occur. Given that the
Internet is more of a business community than a scientific
community, appropriate ambiguity should be openly accepted.

The original 4 NSF sanctioned NAPs should retain a historical
differentiation from all others due to their significance in
transitioning the original NSFNet Internet to privately funded
backbones.

To assert a 'proper use' of the term "NAP" oversteps logic
and implies an arrogantly fascist assertion of perspective.

So I think.

-alan

ps. in a few gogle searches for the paper, I did find a
    reminiscent article on the assumptions and arguments back in
    1994 circa NSFNet cessation of Internet Backbone funding at:

    http://www.cookreport.com/03.07.shtml

10

Something bothers me about this thread, and I think it is the
assertion that there is a proper definition for 'NAP' as
differentiated from 'IXP' or what not.

yup. george santayana was right

11

You mean

"History is always written wrong, and so always needs to be rewritten."

or

"Intolerance is a form of egotism, and to condemn egotism intolerantly is
to share it."

or something else?

12

Bill Woodcock wrote:

About three hundred, by the count that Bill, Antony, and I have been
keeping:

http://www.pch.net/documents/data/exchange-points/ep-in-addrs.txt
http://www.pch.net/documents/data/exchange-points/ep-in-addrs.xls

Australia Adelaide AusBONE AusBONE
  -> There is no Adelaide AusBONE IX; there used to be a
connection
         between AusBone and SAIX, but it is no more.

Australia Adelaide South Australian Internet Exchange
SAIX SE 198.32.240.0/24
Australia Adelaide South Australian Internet Exchange
SAIX SE 203.34.35.0/24
  -> Is the SAIX relocation still in progress or complete?

Australia Brisbane AusBONE-Brisbane Internet Exchange
AusBONE-BIX 198.32.230.0/24
      -> Also uses 198.32.231.0/24 due to the way it's set up.
      -> Location is AAPT, Riverside Centre, 123 Eagle St, Brisbane

Australia Brisbane AusBONE-Brisbane Internet Exchange
AusBONE-BIX 198.32.232.0/24
      -> Duplicate entry with the wrong exchange's IPs (AusBone
Sydney's).

Australia Melbourne AusBONE-Melbourne Internet Exchange
AusBONE-MEL 198.32.234.0/24
  -> Location is AAPT, 530 Collins St, Melbourne

Australia Melbourne Victorian Internet Exchange VIX
      -> IP block is 198.32.194.0/24 (this is a residual exchange from
AUIX,
         and at present the VIX and AusBone-MEL switch fabrics are
linked
         although we intend to separate the switch fabrics Real Soon
Now...)
  -> Location is AAPT, 530 Collins St, Melbourne

Australia Perth Western Australia Internet Exchange WAIX
198.32.212.0/24
  -> Very successful IX, surprised they don't have more IP space
by now :slight_smile:
  -> Location is QV1, 250 St Georges Terrace, Perth

Australia Sydney AusBONE-Sydney Internet Exchange
AusBONE-SYD 198.32.232.0/24
  -> Location is AAPT, 30 Ross St, Glebe

Australia Sydney Pihana Sydney Pihana Sydney
      -> Never heard of them, and there's no real info listed for them

I'm surprised AUSIX isn't in there, or was it already removed after the
Exodus collapse and closure of their Union St, Pyrmont facility where
it was located?

There was also a Paradox Internet Exchange in Sydney (Skygarden, Pitt
St,
Sydney), but we pulled out of there a couple of years back leaving
Paradox to peer with themselves, and they've been through liquidation
at least once since then, so the exchange quite possibly isn't there
anymore :slight_smile:

There were some other IX's - one more in Perth, a few buildings down
from
WAIX (WAIX is in QV1, the other IX was in BankWest, the same building
the
PARNet network is hubbed out of - but PARNet connected to WAIX not the
"other" IX), I'm not sure if it's still there, one in West Sydney (I
forget the name of it) and some guy who set up a 33k6 peering exchange
for small ISPs in Melbourne (I haven't paid much attention to what
happened to that...).

AusBone are attempting to set up some regional IX's in Australia, such
as
Newcastle, Wollongong, etc - none are actually established yet though.

David.

13

I don't believe it has moved at all. There was some talk about it but no action as far as I am aware.

Mark.

14

People should be free to define these terms as they see fit,
with no central authority defining what is what (and I don't
think there should be).

Those first 12 words are the bane of everybody who actually gets work done.

No, regardless of the fact that Americans (myself included) feel the need
to redefine words that had perfectly good and clear meanings in the first
place (myself not included), having a variable and subjective vocabulary
means that succinct, objective discussion is nearly impossible.

There are dozens of papers written on this topic, some dating back over a
thousand years. Scientists have long gotten past this battle. Learn from it.