- NANOG - lists.nanog.org

IX's
by Roeland Meyer 10 Jan '01

10 Jan '01

> From: Vadim Antonov [mailto:avg@kotovnik.com] > Sent: Tuesday, January 09, 2001 3:32 AM > > You mean you really have any other option when you want to > interconnect > few 300 Gbps backbones? :) Both mentioned boxes are in 120Gbps range > fabric capacity-wise. If you think that's enough, i'd like > to point out > at the DSL deployment rate. Basing exchange points at > something which is > already inadequate is a horrific mistake, IMHO. All one has to do is look at PAIX. The whole system looks like it is being used at real close to max capacity. I have a client at AboveNet and my systems are on a CerfNet block. PAIX is between us. I feel their pain. > Exchange points are major choke points, given that 80% or so > of traffic > crosses an IXP or bilaterial private interconnection. > Despite the obvious > advantages of the shared IXPs, the private interconnects between large > backbones were a forced solution, purely for capacity reasons. and they aren't keeping up with the growth. This entire IX thread has been interesting. But, it appears to be one of those "good theory, implementation sux" sort of things.

3 2

Re: RFC1918 addresses to permit in for VPN?
by Sean Donelan 09 Jan '01

09 Jan '01

On Thu, 28 December 2000, "Miguel A.L. Paraz" wrote: > Our DSUA filter had to have a small hole punched in since a customer had a VPN > (I do not know yet as to what kind) which was receiving packets sourced at > 172.17.x.x. Is this a misconfiguration on the sender's end, or a "feature." > > I think there was earlier discussion on VPN's requiring ICMP (echo?) Its not a very private or virtual network if it leaks addresses into a data stream visible to your filters.

21 47

Re: IX's
by Paul Vixie 09 Jan '01

09 Jan '01

rmeyer(a)mhsc.com (Roeland Meyer) writes: > All one has to do is look at PAIX. The whole system looks like it is being > used at real close to max capacity. nope. some paix customers sometimes operate their ports at or near capacity. however, the paix ISO-L2 switch fabric is made up of switches whose backplanes have quite a bit of headroom, and trunks between those switches which have quite a bit of headroom. in spite of this ample amount of headroom, there are a LOT of private network interconnects between paix's customers, which are presumably being used for private peering. the "whole system" is being used at nowhere near its current capacity, and if we ever see a capacity limit on the horizon we will upgrade, upgrade, and then upgrade. > I have a client at AboveNet and my > systems are on a CerfNet block. PAIX is between us. I feel their pain. probably you should contact abovenet and/or cerfnet and find out why that is. it sure as hell isn't because of any capacity limits inside PAIX itself. > > Exchange points are major choke points, given that 80% or so of traffic > > crosses an IXP or bilaterial private interconnection. Despite the obvious > > advantages of the shared IXPs, the private interconnects between large > > backbones were a forced solution, purely for capacity reasons. > > and they aren't keeping up with the growth. in what way? > This entire IX thread has been interesting. But, it appears to be one of > those "good theory, implementation sux" sort of things. a lot of companies live and die according to paix's ability to carry bits or photons or electrons to from other internet companies and/or circuit providers. i like to think that if "implementation sux", i'd've heard more about it before now. but please feel free to educate me. (maybe you should educate me offline and then post a summary back to the list.) -- Paul Vixie <pvixie(a)mmfn.com> CTO and SVP, MFN (NASDAQ: MFNX) Abovenet, PAIX, and MIBH are subsidiaries of Metromedia Fiber Network, Inc.

1 0

RE: net.terrorism
by Roeland Meyer 09 Jan '01

09 Jan '01

> From: Randy Bush [mailto:randy@psg.com] > Sent: Tuesday, January 09, 2001 11:36 AM > > >> ethical question like this. > > I was under the misconception the 'O' in NANOG stood for > > 'Operations' not 'Ethics'. But this would explain lots > > of other posts here which I had previously considered > > off topic. > > s/operations/orations/? That can't be; whilst I've read an awful lot, I haven't "heard" a thing <g>. Is there an aural back-channel to NANOG? <going back to configuring root servers now ...>

1 0

Re: net.terrorism
by Jerry Scharf 09 Jan '01

09 Jan '01

I can't resist trying to sum this up: Somebody wants to talk to an ORBS server and has Abovenet transit. They don't like being caught in the orbs/abovenet mess. So rather than a) getting a new provider b) getting another provider and fixing the local routing to use them to ORBS the person opted for c) send a misleading flame message to NANOG Operational issues of enforcing trafic delivery policies for customers has been discussed till it's dead. Everyone knows where they stand, and anyone who does it has is as part of their service agreement. As has been pointed out, this is not a subject that will engender any sympathy in this group. does anything else need to be said? If not, let's let this die. jerry

1 0

Original peering policies (com-priv archives?)
by Sean Donelan 09 Jan '01

09 Jan '01

Has anyone kept an archive of the various peering policies different providers had over the years? For example, Sean Doran published Sprint's peering policy on COM-PRIV a number of years ago. And I believe the original peering policy on the CIX router included an "AUP-Free" clause, that no provider would impose an acceptable use policy on traffic between the providers. Or how could I forget ANS's original CO+RE settlment plan (i.e. everyone pays ANS). But I'm just going from memory, and some of those brain cells aren't as young as they were. I'm interested in seeing a timeline and evolution of policies over the last 10 years. I know some have changed.

4 3

Re: net.terrorism
by Paul A Vixie 09 Jan '01

09 Jan '01

this is not relevant to nanog, and because of the high likelihood that any private reply i send will end up being posted to nanog later, i'm going to drop out of this thread after this reply. > What I would expect is that you would choose between two things: > > 1. you blackhole but do NOT announce those netblocks; > 2. you annonce AND deliver traffic to every host in it; > > Don't you agree that announcing means delivering traffic? no. > Especially for customers. you should read the terms of service in your abovenet customer agreement to find out what customers can expect from abovenet. > > why are we discussing this on nanog? > > Because Above.net seems violates the first thing needed in > internetworking: trust. If you tell me you will deliver traffic to $blah, > I think I may expect you to do so. That's my whole point. Nullroute as > much as you want but don't announce it on your border routers... oh good, yet another pointless debate held in yet another ivory tower. bye.

1 0

UUNET peering policy
by Barrows, Jeff 09 Jan '01

09 Jan '01

UUNET has just posted their new peering policy: http://www.uu.net/peering/ ...even if you don't like it, you have to give them credit for finally making it available to the public without requiring an NDA. ;-> - jsb

8 7

Re: UUNET peering policy
by smd＠clock.org 09 Jan '01

09 Jan '01

| How many medium sized ISPs do we need to get to work together to make a | "meta ISP" and get peering with UUNet ;) 0, because there is no single legal entity owning the facilities in each of those 15 states. Alternatively, 0, because it is not in UUNET's business interests to give a 100% discount to some set of their target market. However, a big enough buyer is likely to be able to negotiate something approaching a discount of 100% on a paid-peering-like service, so you might phrase your question, "how much discount would a meta ISP of size X be able to negotiate on a single fixed-term contract?". As Sean Donelan noted, there is a difference between being qualified to negotiate a no-settlement peering agreement, and actually entering into one. The policy is designed to make it very easy to answer "no, you don't qualify" without having to enter into long business discussions under NDA, which is time-consuming. It also answers the question of, "does your no today mean no tomorrow also?" Sean.

1 0

Re: UUNET peering policy
by Jeff.Hodges＠kingsmountain.com 08 Jan '01

08 Jan '01

Folks here may find the below interesting. also available here.. http://www.interesting-people.org/200101/0015.html JeffH ----- Date: Mon, 08 Jan 2001 16:01:33 -0500 To: ip-sub-1(a)majordomo.pobox.com From: Dave Farber <farber(a)cis.upenn.edu> Subject: IP: A watershed event has occurred with no fanfare... Mime-Version: 1.0 Sender: owner-ip-sub-1(a)admin.listbox.com Precedence: list Reply-To: farber(a)cis.upenn.edu Content-Type: text/plain; charset="us-ascii"; format=flowed Content-Length: 2706 >Date: Mon, 08 Jan 2001 13:41:14 -0700 >From: Rodney Joffe <rjoffe(a)centergate.com> >To: Dave Farber <farber(a)cis.upenn.edu> >Subject: A watershed event has occurred with no fanfare... > >Hello Dave, > >As you probably know from your time at the FCC, as well as your earlier >'net history, probably the most controversial and secretive subject >relating the the true backbone, and infrastructure of the Internet, has >been that of peering and peering relationships. For clarification, >peering in the Internet sense is the exchange of traffic between two >networks without filtering, limitation, or fee. > >As the founder of the original Genuity, in 1993-4 I was the last of the >networks to be party to the multi-lateral peering agreement that made >the commercial Internet work. From that moment on, it became virtually >impossible for anyone else to connect and exchange traffic with the rest >of the Internet without paying a transit or traffic fee to someone else. >The whole subject became mired in obfuscation and secrecy, and over the >next 7 years, the company that everyone needed to peer with but no new >company could was UUNet (now Worldcom/MCI/UUNet). Even with the other >major networks at the time (Sprint, MCI, BBN, AT&T, IBM etc.) this was >difficult, because the secret sauce seemed to be finding out what a >given network's peering requirement was, and then meeting those >requirements. And no major network ever published it's requirement. > >The requirement was generally political, but articulated in some vague >technical specification (e.g. to be present at 5 public exchange points, >with a point-to-point clear channel T3 between all points, and at least >one redundant path between any two exchange points, etc.). > >So the rest of the Internet industry basically gave up on ever being >able to join the chosen few who peered at the center of the Internet >without paying any form of settlement fees. > >Then, apparently in the last few days, a clear statement on peering >policy and requirements has appeared on the UUNet website - see >http://www.uu.net/peering/ This is a remarkable document in that for the >very first time ever, in a public form, it sets standards for >settlement-free (i.e. no charge) peering with UUnet, the world's largest >Internet network. > >If I understand the document correctly, anyone who meets their clear >requirements will be able to exchange traffic with them at no charge. > >This will undoubtedly change the landscape of the Internet. > >Regards, >-- >Rodney Joffe >CenterGate Research Group, LLC. >http://www.centergate.com >"Technology so advanced, even we don't understand it!"(SM) For archives see: http://www.interesting-people.org/ ------- End of Forwarded Message

1 0