October 1996 - Test - lists.nanog.org

Re: Intra/Inter - was Inet-II
by Paul Ferguson 14 Oct '96

14 Oct '96

I forgot to include my original smiley. ;-) - paul At 11:48 AM 10/14/96 -0400, Curtis Villamizar wrote: >> Another suggestion? Lacking one, one might suggest that the complexity >> of the 'mesh' needs to be simplified. >> >> - paul > > >Paul, > >It's that sloppy multiprovider model. Let's go back to one NSFNET >core. :-) > >Curtis > >ps- For the *extremely* humor impaired -- I'm just joking. I'm not >seriously suggesting that. I am suggesting to Paul through (friendly) >sarcasm that we have to live with routing complexity. >

1 0

Re: Intra/Inter - was Inet-II
by Havard.Eidnes＠runit.sintef.no 14 Oct '96

14 Oct '96

> ftp://www.isi.edu/div7/ra/Publications/bgp_osc.ps.gz This one should be ftp://ftp.isi.edu/ra/Publications/bgp_osc.ps.gz - Havard

1 0

Re: generators, etc....
by Steve Norton 14 Oct '96

14 Oct '96

At 11:56 AM 10/14/96 -0400, Curtis Villamizar wrote: >If we hold BBN to the same standards that the telco industry uses to >come up with 99.999%, BBN was up but "a few customers" experienced a >localized outage. Discussion of reliability standards has gotten a little vague. To throw some facts into the confusion: I once managed to beat our Telco reps hard enough for them to provide actual numbers. This was after a 4 hour outage on an OC-3 ring that "could self heal in 50 milliseconds." Ameritech provided the usual mushroom treatment, but MFS actually quoted their Operations procedure AP2402, which says: For a DS3 facility 0-50 miles: Availability on a 12 month period is 99.997% Error free seconds > 99.992% over 24 hour period (7 seconds/day) Bit error rate < 2.5*10E-11 "Performance tests shall be conducted at DS-3 level for 72 hours." Apparently this is copied from AT&T PUB 62411. Cisco gives me the mushroom treatment, but assuming they shoot for something like 99.992% uptime, a simple DS3 Internet connection could have no better than 99.98% reliability. Does anyone even come close to this number? With all of the ATM deployed recently, I cant imagine any ISP even comes close. Steve Norton 312-496-4601 VP Operations 312-496-4499 (fax) InterAccess Co steve(a)noc.interaccess.com

2 1

Re: Intra/Inter - was Inet-II
by Paul Ferguson 14 Oct '96

14 Oct '96

Herr Manning, At 11:15 AM 10/11/96 -0700, bmanning(a)ISI.EDU wrote: >> >> You mean that routing protocols need to take account of policies? ;-) >> >> Or do you mean that BGP isn't good enough? >> >> Michael Dillon - ISP & Internet Consulting > >Yes... (silly :) >Yes... (serious) > >BGP has a number of serious flaws. The more complex the routing mesh, the >more they will become visable. And if what I see is correct, IDRP, the >BGP heir, has the same sets of problems. > Another suggestion? Lacking one, one might suggest that the complexity of the 'mesh' needs to be simplified. - paul

4 5

Re: Internet II is coming...
by salo＠msc.edu 14 Oct '96

14 Oct '96

> Subject: Re: Internet II is coming... > Date: Wed, 09 Oct 1996 17:20:17 -0400 > From: Fletcher E Kittredge <fkittred(a)biddeford.com> > > On Wed, 9 Oct 1996 12:13:03 -0700 (PDT) Michael Dillon wrote: > > My ISP has a 10Mbps fibre ATM circuit to BCNet in Vancouver that was > > installed in Aplril 1995. From there it connects to CA*Net which has a T3 > > into MCI Seattle as well as links to the East where more T3's head south > > to MCI. The T3's were T1's until Sept 1995 at which point there was a > > small improvement in speed at times. > > So a 10Mbps fiber ATM circuit translates, after the cell tax and > packet stredding, to 6Mbps on a real link. Is this not slow enough to > be a bottleneck? And, an OC-12c ATM circuit can provide nearly 540 Mbps of TCP data, after SONET, ATM, IP and TCP overheads. Perhaps, I missed your point... -tjs

2 1

CIX Filtering...
by salo＠msc.edu 14 Oct '96

14 Oct '96

Does anyone remember when the CIX proposed to filter networks of customers of CIX members which were not themselves CIX members? Was it as early as 1995? (We are trying to review a paper on Internet exchanges. It seems ironic that this effort by the CIX to restrict traffic was rendered ineffective by the same technique the CIX used to overcome the impediments created by the NSFNET AUP, namely alternative paths for "unacceptable" traffic.) Thanks, -tjs

5 7

Re: generators, etc ...
by Sean Donelan 14 Oct '96

14 Oct '96

In article <m0vCc8h-001K9SC(a)on-ramp.ior.com>, davids(a)on-ramp.ior.COM (David J. Schmidt) writes: > An editorial in Unix Review (October issue I believe) talked about > some outages that their site had suffered and mentioned other > outages. Then the writer asked the question "Who's at fault?". His > answer was that the consumer was at fault, because the consumer is > unwilling to pay the rates necessary to pay for the level of service > that they demand. > > We all scream for fixes when these outages occur. How much are we > (and our customers) willing to pay for them? Heavy sigh. Even if you are willing to pay more you won't find anything better to buy. I've been trying to buy better service for years, and haven't found a seller willing and able to commit to deliver anything better. Multi-homing to multiple naps, leasing separate lines, maintaining a 24 hour NOC, and general management of the whole thing is expensive and time consuming. I would love to outsource the whole mess and go back to making money on our real product. There are customers out there willing to pay more, because they are already paying a lot more. They are fairly easy to spot because they are doing lots of crazy things like multi-homing, leasing their own lines, buying big cisco's and connecting them to NAPs. Show me you can deliver. -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

1 0

Re: BBN outage
by Apu 14 Oct '96

14 Oct '96

On Fri, 11 Oct 1996, Charles R. Hoynowski wrote: > We're trying to figure out how long BBN's been down, are you talking > 12:30am PDT? If so, then they been down 16 hours. I started seeing mail > failures to Stanford (a big BBN site) at 9:33am PDT and first got email > about the problem from another ISP at around 11am. I haven't seen anybody answer this, and while I suppose everybody could call the BBN NOC and ask, that's probably not a useful way for them to spend their time. Therefore, for those who don't know but want to, and for the record (archive)... According to BBN ticket # 60414, the problem started at 03:27 UTC (October 11, 1996). As of October 12, 1996 01:54 UTC (the latest entry on the ticket), BBN believes all affected customers should be back online. This is the last entry on the ticket as of now. _ | | |-| | |pu <apu(a)spfld.com>

1 0

Re: CIX Filtering...
by salo＠msc.edu 14 Oct '96

14 Oct '96

Thanks again, -tjs

1 0

Re: CIX Filtering...
by salo＠msc.edu 14 Oct '96

14 Oct '96

Thanks, -tjs

1 0