- NANOG - lists.nanog.org

Re: GigaRouter (Was Re: Cisco as Big Brother))
by Peter Lothberg 29 Oct '96

29 Oct '96

>Ok, now we are building a OC3 and DS3 backbone and a PC can't come close >to the PPS a cisco or NetStar can. My only question is should I go with >the NetStar/Cascade or just cisco network? > >Nathan Stratton CEO, NetRail, Inc. Tracking the future today! A high PPS rate is not the goal here, it's the system performance of the total network that counts. Some boxes with high PPS numbers have no buffers, so they get outperformed by a PC, even thus the PC can't push that many packets. I would not be surprised if a dual Pentium box outperforms one of the most common routers.... -Peter

1 0

RE: Airplane crashing into Atlanta-NAP
by Chris A. Icide 28 Oct '96

28 Oct '96

Michael Dillon brilliantly wrote.... > > The only joke here is that some people think that the Internet exchange in > a city the size and importance of Atlanta can be adequately housed on the > 5th story of an office building. > > I could understand it if Atlanta had 20 or 30 exchange points, but when > there is only one exchange the city deserves better than a mickey-mouse > 5th floor operation. > Oh, I agree wholeheartedly!!!!! And who would ever imagine that an exchange point in the basement of a mere parking garage would have achieved the level of traffic it has! Now that I've thrown in my share of late night sarcasm, It would interest me greatly to understand exactly why you came to the following conclusions: 1. The "operation" in Atlanta is Mickey Mouse 2. The floor # of a bulding affects the quality of the Exchange Point 3. The type of building affect the quality of the Exchange Point 4. A city the size of Atlanta needs more than 1 Exchange Point Truly I would be very interested in your thoughts on these items, as well may a few other folks on this list. Chris A. Icide Nap.Net, L.L.C.

9 11

Humor (was Airplane crashing into Atlanta-NAP)
by William Allen Simpson 28 Oct '96

28 Oct '96

> From: Avi Freedman <freedman(a)netaxs.com> > Could we stop this, please? > Actually, I enjoyed the recent humor! > Either Netrail's NAP will be useful or it won't be. > It'll be useful as a local exchange if local providers go. > It'll be useful as a place to sell transit of transit providers go. > But as to the rest of the thread, yes, it's pretty well exhausted. Except, I hope that the locals will join together and make this MIX/NAP work, or if it is overpriced or otherwise problematic, join together to make _something_ work.... > It'll be useful as a place to offload huge chunks of Internet traffic > between larger providers only if they go, which they don't seem to > be interested in yet. > Jointly buy _transit_ service in a bigger pipe, and specify that you will only buy it at the NAP ;-) > > ... > > If anyone is interested in setting up an alternative Atlanta NAP, I have > > a very spacious first-floor apartment which I am more than happy to > > volunteer to the effort. It's even somewhat sunk below ground level, and > > I think my neighbors wouldn't mind the occassional running of a generator. > > > > I've already got 10MB/s, which puts the Northside-Circle-NAP on par with > > MAE-NY, right? > > ... > > Todd Graham Lewis Linux! Core Engineering > > Mindspring Enterprises tlewis(a)mindspring.com (800) 719 4664, x2804 > > WSimpson(a)UMich.edu Key fingerprint = 17 40 5E 67 15 6F 31 26 DD 0D B9 9B 6A 15 2C 32 BSimpson(a)MorningStar.com Key fingerprint = 2E 07 23 03 C5 62 70 D3 59 B1 4F 5E 1D C2 C1 A2

1 0

Re: Provider credibility - does it matter? was Re: Inter-provider relations
by Bill Cessna 28 Oct '96

28 Oct '96

>>From: jakob(a)jubii.dk (Jakob Faarvang) >>To: "Alex.Bligh" <amb(a)xara.net>, roll(a)stupi.se, nanog(a)merit.net >>Subject: Re: Provider credibility - does it matter? was Re: Inter-provider >relations >>Date: Sun, 27 Oct 1996 14:42:43 GMT >>Organization: cybernet.dk >>References: <199610271111.LAA11077(a)diamond.xara.net> >>Sender: owner-nanog(a)merit.edu >> >>On Sun, 27 Oct 1996 11:11:10 +0000, you wrote: >> >>>> >>>> I don't think this is correct. Ebone peers with other pan-european >>>> networks of comparable size, IBM, BT and EUnet comes to my mind. >>>> >>>In fact, last time I heard, EBONE and EUnet's peering policy >>>bore an uncanny resemblence to those of Sprint & MCI (albeit >>>bandwidth scaled). >> >>From Ebone's own pages, BT is listed as a customer to Ebone, and it is >>said the IBM and EUnet have transit agreements with Ebone. I can't >>find any direct text stating Ebone's peering policies, but on the >>Amsterdam IX page (http://www.ams-ix.net/) the following is said: >> >>Policy: Does generally not peer except with other larger ISPs after >> negotions. Become a member and we'll peer with you! >> >>- Jakob Faarvang >> >> IBM Global Network peers with EBONE. We do not get or buy transit from EBONE. The EBONE home pages (last time I checked) did not provide enough information to crisply tell who were members and who were peering partners. To peer, you must connect with EBONE at three places at E-1 speed as a minumum. Web pages for EBONE are at (http://www.ebone.net/) and say this if you read carefully. (I think they do...) Members get transit. That is what being a member of EBONE is about and how the EBONE backbone provides value to its members (members can be thought of in terms of a customer). It is a model that works well in Europe and provides for all ISP's to join. Fees that support the EBONE backbone are based on the members access to the EBONE backbone line speed. Bill Cessna Bill Cessna, IBM Global Network, Network Architecture Internet: wcessna(a)vnet.ibm.com Internal: IPNET(WCESSNA) Voice: +1-813-878-3947 Fax: +1-813-878-4250

1 0

US transit
by Per Gregers Bilse 28 Oct '96

28 Oct '96

I'd like to set one thing straight: It is completely bogus if people run around believing there are huge amounts of non-US traffic transiting the US, clogging up exchanges and whatnot. The volume of this kind of traffic is completely in the noise compared to traffic which is sourced and/or destined in/for the US; any argument suggesting the opposite is misguided or wilfully misleading. -- ------ ___ --- Per G. Bilse, Mgr Network Operations ----- / / / __ ___ _/_ ---- EUnet Communications Services B.V. ---- /--- / / / / /__/ / ----- Singel 540, 1017 AZ Amsterdam, NL --- /___ /__/ / / /__ / ------ tel: +31 20 5305333, fax: +31 20 6224657 --- ------- 24hr emergency number: +31 20 421 0865 --- Connecting Europe since AS286 --- http://www.EU.net e-mail: bilse(a)EU.net

1 0

Re: Provider credibility - does it matter? was Re: Inter-provider relations
by scope＠cais.cais.com 28 Oct '96

28 Oct '96

At 06:28 PM 10/28/96, Jim Dixon wrote: >On Mon, 28 Oct 1996, Jian Li wrote: > >> > The point is that GSL's peering policies are radically different from >> > those of SprintLink, whereas Cable & Wireless seems to have the same >> > policies worldwide. Similarly, I know that GSL's prices for transit >> > used to be much higher than SprintLink's and I suppose still are. >> >> As what I know, GSL's peering policies have two parts, international and >> US domestic. The US domestic part is exactly the copy of SprintLink. >> International part is based on the idea of traffic localization. > >Fine. This still leaves Sprint with two very different policies, one >for the US and one for the rest of the world, whereas CWIX apparently >has one policy for the whole world -- or at least for North America and >Europe. I believe that over the next few years this is how things are >going to go: peering policies and prices for transit will converge, at >least in these two large markets. Sprint has appearently made the decision to treat Sprintlink and GSL as two entirely seperate networks. This makes sense based on the way that they structured Global One. The GSL product is appearently intended to be an international Internet product. In this sense they really look more like a no U.S. provider with the rate structure and cost basis to match. If the idea is to drag the traffic to the U.S. (whether because it is cheaper to do so, U.S. traffic is a bigger draw, or the U.S. telephony infrastructure and price structure causes the U.S. to be attractive for Asia - Europe transit) then GSL has to price accordingly. Sprintlink, being primarily focused on the U.S. domestic market has a much easier time with underlying costs and customer concentration. Their terrestrial fiber (as Peter L. pointed out) makes it so. > >Americans tend to forget that Europe has a larger population and more >money than the US; these basic facts are not going to let the current >large imbalances persist. The imbalances will fade when and if Europe takes its huge population and bigger bucks (marks, pounds,francs,ecus) and wields them like one economy, not a dozen. What drove the U.S.centric model in the first place wasn't lack of European peering, it was that, way back when, It was cheaper to get to Paris (or name your city in Europe) via Washington than straight on. j >-- >Jim Dixon VBCnet GB Ltd +44 117 929 1316 fax +44 117 927 2015 >http://www.uk.vbc.net VBCnet West +1 408 971 2682 fax +1 408 971 2684 > > Jeff Studds Internet Stuff scope(a)cais.com phone +1 301 699 1840 The internet: It's not just an adventure... it's a job:-)

1 0

Cisco Memory.
by plynch＠jefferson.ind.net 28 Oct '96

28 Oct '96

Nathan, I have to follow up on this comment you made. I work for an ISP in Indianapolis called IHETS. We have 7 cisco 7500 series routers on our backbone. Wre just spent a chunck of money to upgrade the RAM on them to 64Meg. If yoiu are saying that regular SIMMS work in Cicso routers, then I'll shoot myself in the foot. I'd like to know if you have used this method and how well it works out in the field. Does latency increase with *unqualified* Cisco memory, or is the *unqualified* *qualified* Cisco ram stuff just a big ruse? Thanks for your input. > > On Mon, 21 Oct 1996, Alexis Rosen wrote: > > > alex(a)relcom.eu.net writes: > > > and WHY have I to pay 3,500$ for the 32Mb ram if this RAM costs > > > 600$ on the free market???). > > > > Dunno for sure about the 4xxx, but the 7xxx use standard SIMMs, and Cisco > > has even qualified some manufacturers' parts. Oddly enough, they have > > also *disqualified* some parts that you'd think were exactly the same. I > > don't know the details there. The sales types obviously don't like to > > talk about it, but you really *can* get $12k of 7500 memory for $800. And > > they'll even agree that it works, if you talk to the right cisco folks. > > Yep, it uses standard SIMMs, I have upgraded many of them.

13 20

Re: Provider credibility - does it matter? was Re: Inter-provider relations
by azeem＠dial.pipex.com 28 Oct '96

28 Oct '96

>Karl Denninger <karl(a)Mcs.Net> wrote: > >>Any provider that does not recognize the value of bilateral, no-settlement >>peering anywhere that its cost-effective for both parties (ie: if you have >>traffic destined for me, get it on MY network where I'm being paid to >>carry it and let ME figure the rest out!) deserves what they get. > >Zero-settlement peerings open to anyone are demonstrably amount to >subsidies from large peers to small. Absolutely correct: can I direct you to two of our articles on the subject (and a useful graphic): http://www.economist.com/issue/19-10-96/ld4401.html http://www.economist.com/issue/19-10-96/sf0774.html http://www.economist.com/issue/19-10-96/sf1.gif >Zero-settlements work only when peers are of comparable size. >Any attempt to extort pressure to force it upon anyone simply >causes large folks to flee. That's true: in a conversation with Phil Lawlor (CEO of AGIS), he told me that one of his customers ISPs, who was also peering with AGIS at the Mae-E, indulged in "bandwidth stealing" by receiving their Usenet news feed across Mae-E rather than as an AGIS customer across the few T-1s they had to AGIS. Surely a hierarchy which implements the rule you outline above works: a pair of small regional ISPs may want to peer regionally for good reason, because they are the same size there isn't much inequity in this relationship. Two large ISPs may wish to do the same. However, a small ISP who wants its traffic carried by a large ISP (i.e. peer with them) may have little choice but to pay for this traffic to be carried: so becomes a customer. That way you have a hierarchy as follows: Bacbkbone providers who may peer with each other Large regional providers who are customers of the backbone providers (so by DS-3s or whatever from BBNplanet, UUnet) and who may peer regionally. While also present at bix exchanges, they are unlikely to peer with the big six/seven/eight but may well peer with foreign ISPs at those exchanges. Azeem Azeem Azhar vx: 0171-830 7133 The Economist fx: 0171-839 2968 25 St James's Street London SW1A 1HG

14 39

Re: October 1996 Meeting Notes
by sob＠academ.com 28 Oct '96

28 Oct '96

That's because I said the notes "would be appearing" this week. They have not appeared yet. -- Stan | Academ Consulting Services |internet: sob(a)academ.com Olan | For more info on academ, see this |uucp: {mcsun|amdahl}!academ!sob Barber | URL- http://www.academ.com/academ |Opinions expressed are only mine.

1 0

October 1996 Meeting Notes
by sob＠academ.com 28 Oct '96

28 Oct '96

They will start appearing this week at this URL: http://www.academ.com/nanog/oct1996 Enjoy. Comments welcome.

2 1