- NANOG - lists.nanog.org

Network Perf. Speaker Needed
by Susan R. Harris 08 Apr '97

08 Apr '97

Greetings - we're starting initial agenda planning for the next NANOG meeting, to be held June 5-6 in Tampa. Could we interest someone out there in giving a 20-to-30 minute "Introduction to Network Performance Measurement?" The talk would explain what it's possible to measure and give examples of useful types of internal and external measurements (e.g., link utilization and DS3/DS1 MIB stats for internal metrics, one-way delay and packet loss for external.) What we're aiming for is a discussion of performance measurement fundamentals - not specifics on what you're measuring in your network. The presentation would be followed by a panel discussion with short reports on specific tools (SNMP tools, ISODE, CMU, tcl etc.), to be moderated by Randy Bush. We'd also like volunteers to give short reports as part of the panel. Please let Bill Norton or me know if you're interested in helping out. And while we're at it, are there other topics you'd like to hear about at the meeting? ------------------------------------------------------------------------ Susan R. Harris, Ph.D. Merit Network, Inc. srh(a)merit.edu Phone: (313) 936-2100 Fax: (313) 747-3185 ------------------------------------------------------------------------

1 0

Problems with the inet-acess list
by Michael Dillon 08 Apr '97

08 Apr '97

I contacted someone at BSDI who tells me that Tony Sanders knows about the list problem and is working on it. Michael Dillon - Internet & ISP Consulting Memra Software Inc. - Fax: +1-250-546-3049 http://www.memra.com - E-mail: michael(a)memra.com

1 0

off-topic backhaul question
by andrew khoo 07 Apr '97

07 Apr '97

this is possibly off-topic, so i apologize in advance. we require to find suitable contacts to assist us in: 1) arranging a DS3 ATM connection into the Pacbell ATM cloud from PAIX 2) arranging a DS3 ATM connection from 49 Geary, San Francisco into the Pacbell ATM cloud. anyone can assist? --- andrew khoo plumber @ OGN ph: +61-2-9223-2777 gsm: +61-416-161-234

1 0

RE: Peering points
by Dave Van Allen 07 Apr '97

07 Apr '97

FYI Sprint is not allowing anything larger than 6Mb from Pensauken, circa April 4, 1997, for customer connections that is. Best regards, David Van Allen - You Tools Corporation / FASTNET(tm) dave(a)fast.net (610) 289-1100 http://www.fast.net FASTNET - PA/NJ/DE Internet Solutions >-----Original Message----- >From: Avi Freedman [SMTP:freedman@netaxs.com] >Sent: Saturday, April 05, 1997 5:03 AM >To: SEAN(a)SDG.DRA.COM >Cc: nanog(a)merit.edu >Subject: Re: Peering points > >> Or is the entire thing irrelevant, because everyone is moving to private >> bilateral connections. And the important thing is how many providers >> does someone peer, not how many exchange points they're connected to. >> For the price of one Sprint-NAP connection, I can get several connections >> to Canada. And we have a lot more customers in Canada than at the >> Sprint-NAP. We already peer with everyone at the Sprint-NAP, or been >> turned down by them elsewhere. So one more exchange point doesn't buy >> much. >> -- >> Sean Donelan, Data Research Associates, Inc, St. Louis, MO >> Affiliation given for identification not representation > >The SprintNAP is a much less congested place than MAE-East (though most >providers also have less capacity out of it [besides Sprintlink, GSL, >ICM, etc...])... > >Avi >

7 7

RE: Peering points
by Steve Schnell, Sprint Corporation 07 Apr '97

07 Apr '97

David, et al: You must understand that SprintLink and the NAP, which is run by Sprint Government Systems Division, are totally separate organizations. Actually, SprintLink is a customer of the NAP - just like all other ISPs there. It is a policy of GSD to limit its NAP customers to DS3 and higher rates. We don't support T1s because we feel that speed is inappropriate for peering with DS3 ISPs. We have migrated all our customers onto the GIGAswitches and we no longer support FDDI concentration. Terminating a T1 customer to a NAP GIGAswitch would be inefficient use of a switch port and may contribute to the so-called head-of-line blocking phenomena. Also, T1s are a headache to wire and support. They break far more frequently than DS3s. Steve At 19:46 -0400 4.6.97, Dave Van Allen wrote: >I'm not trying to imply anything, just that there is a current limit on >customer connections there. Your routing through Pennsauken seems to be >pretty event free, so maybe Sprint simply wants the NAP to stay optimum >and not overloaded. This would be a Good Thing. > >Best regards, > >David Van Allen - You Tools Corporation / FASTNET(tm) >dave(a)fast.net (610) 289-1100 http://www.fast.net >FASTNET - PA/NJ/DE Internet Solutions > >>-----Original Message----- >>From: Avi Freedman [SMTP:freedman@netaxs.com] >>Sent: Sunday, April 06, 1997 7:27 PM >>To: Dave Van Allen >>Cc: freedman(a)netaxs.com; SEAN(a)SDG.DRA.COM; nanog(a)merit.edu >>Subject: Re: Peering points >> >>Interesting... >>We've seen the SprintNAP as a good place to send data to Sprintlink >>and MCI - and UUNET, when we were peering with them there. >> >>Avi >> >>> FYI Sprint is not allowing anything larger than 6Mb from Pensauken, >>> circa April 4, 1997, for customer connections that is. >>> >>> Best regards, >>> >>> David Van Allen - You Tools Corporation / FASTNET(tm) >>> dave(a)fast.net (610) 289-1100 http://www.fast.net >>> FASTNET - PA/NJ/DE Internet Solutions >>> >>> >-----Original Message----- >>> >From: Avi Freedman [SMTP:freedman@netaxs.com] >>> >Sent: Saturday, April 05, 1997 5:03 AM >>> >To: SEAN(a)SDG.DRA.COM >>> >Cc: nanog(a)merit.edu >>> >Subject: Re: Peering points >>> > >>> >> Or is the entire thing irrelevant, because everyone is moving to private >>> >> bilateral connections. And the important thing is how many providers >>> >> does someone peer, not how many exchange points they're connected to. >>> >> For the price of one Sprint-NAP connection, I can get several >>>connections >>> >> to Canada. And we have a lot more customers in Canada than at the >>> >> Sprint-NAP. We already peer with everyone at the Sprint-NAP, or been >>> >> turned down by them elsewhere. So one more exchange point doesn't buy >>> >> much. >>> >> -- >>> >> Sean Donelan, Data Research Associates, Inc, St. Louis, MO >>> >> Affiliation given for identification not representation >>> > >>> >The SprintNAP is a much less congested place than MAE-East (though most >>> >providers also have less capacity out of it [besides Sprintlink, GSL, >>> >ICM, etc...])... >>> > >>> >Avi >>> > >>> >> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Phone: 1.816.854.2113 Fax: 1.816.854.2201 Sprint Pager: 1.800.724.3329, PIN 398.6644 Outside U.S.: +1.619.279.8495, PIN 398.6644 Text Page via Internet: 3986644(a)pagenet.net or via WWW at: http://www.sprint.com/home/custserv/pager/sendpage.html * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

1 0

RE: Peering points
by Dave Van Allen 06 Apr '97

06 Apr '97

I'm not trying to imply anything, just that there is a current limit on customer connections there. Your routing through Pennsauken seems to be pretty event free, so maybe Sprint simply wants the NAP to stay optimum and not overloaded. This would be a Good Thing. Best regards, David Van Allen - You Tools Corporation / FASTNET(tm) dave(a)fast.net (610) 289-1100 http://www.fast.net FASTNET - PA/NJ/DE Internet Solutions >-----Original Message----- >From: Avi Freedman [SMTP:freedman@netaxs.com] >Sent: Sunday, April 06, 1997 7:27 PM >To: Dave Van Allen >Cc: freedman(a)netaxs.com; SEAN(a)SDG.DRA.COM; nanog(a)merit.edu >Subject: Re: Peering points > >Interesting... >We've seen the SprintNAP as a good place to send data to Sprintlink >and MCI - and UUNET, when we were peering with them there. > >Avi > >> FYI Sprint is not allowing anything larger than 6Mb from Pensauken, >> circa April 4, 1997, for customer connections that is. >> >> Best regards, >> >> David Van Allen - You Tools Corporation / FASTNET(tm) >> dave(a)fast.net (610) 289-1100 http://www.fast.net >> FASTNET - PA/NJ/DE Internet Solutions >> >> >-----Original Message----- >> >From: Avi Freedman [SMTP:freedman@netaxs.com] >> >Sent: Saturday, April 05, 1997 5:03 AM >> >To: SEAN(a)SDG.DRA.COM >> >Cc: nanog(a)merit.edu >> >Subject: Re: Peering points >> > >> >> Or is the entire thing irrelevant, because everyone is moving to private >> >> bilateral connections. And the important thing is how many providers >> >> does someone peer, not how many exchange points they're connected to. >> >> For the price of one Sprint-NAP connection, I can get several >>connections >> >> to Canada. And we have a lot more customers in Canada than at the >> >> Sprint-NAP. We already peer with everyone at the Sprint-NAP, or been >> >> turned down by them elsewhere. So one more exchange point doesn't buy >> >> much. >> >> -- >> >> Sean Donelan, Data Research Associates, Inc, St. Louis, MO >> >> Affiliation given for identification not representation >> > >> >The SprintNAP is a much less congested place than MAE-East (though most >> >providers also have less capacity out of it [besides Sprintlink, GSL, >> >ICM, etc...])... >> > >> >Avi >> > >> >

2 1

RE: Peering points
by Rodney Joffe 06 Apr '97

06 Apr '97

-----Original Message----- From: Sean Donelan [SMTP:SEAN@SDG.DRA.COM] Sent: Saturday, April 05, 1997 4:50 PM To: nanog(a)merit.edu Subject: Re: Peering points <snip> Heck, I wonder how much congestion could be allievated at MAE-East by people just exchanging traffic via all the places they connect in common. I peer with several providers only at MAE-East, who also have connections elsewhere. But they've never turned up peering at the other places. So all the traffic ends up flowing through MAE-East. Sean, It seems to me that you're being screwed..... As far as I remember, every peering agreement we have (even with Sprint and MCI) says that we agree to peer with each other at *_every_* location that we are both present at. Isn't that the right way to do it? Selective peering is almost as bad as prepending AS's Rodney Joffe Chief Plumber Genuity Inc., a Bechtel company http://www.genuity.net

1 0

Re: Peering points
by Sean Donelan 05 Apr '97

05 Apr '97

>The SprintNAP is a much less congested place than MAE-East (though most >providers also have less capacity out of it [besides Sprintlink, GSL, >ICM, etc...])... The Chicago NAP is even less congested (though most providers have less capacity out of it [besides AGIS, etc]) :-). If you refer back to my exchange point cross-matrix, you'll notice the Sprint NAP doesn't have a single unique connection. The Sprint column under the 1 connection section is empty. That tells me I would find more unique or interesting peering possibilities elsewhere. Or, to say it another way, everyone at the Sprint NAP is connected to at least one other exchange point and already has the option to peer with me (or not). It seems rather pointless to make the effort to connect somewhere, just to be told by the same group of people, no again. Heck, I wonder how much congestion could be allievated at MAE-East by people just exchanging traffic via all the places they connect in common. I peer with several providers only at MAE-East, who also have connections elsewhere. But they've never turned up peering at the other places. So all the traffic ends up flowing through MAE-East. -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

1 0

Re: Peering points
by Sean Donelan 05 Apr '97

05 Apr '97

>You need to update your list BAD. We are connected to all but one of your >NAPs listed. We wanted to be at Sprint-NAP 4 months ago, but they did not >have the gigaswitch port, or the physical space for our hardware. Please do >not put information about other provider up without checking it out first. I guess everyone can see why CERFNET stopped updating their web page about exchange point connections. Sorry, I missed your pacbell nap connection. It appeared down when I pinged it earlier, but now it appears up. I'll update my spreadsheet. I figure people wouldn't be shy about correcting me. I've also noticed they haven't been as forthcoming when I listed a connection that isn't 100% operational. Actually, I'm doing well so far. The correction responses break down like this, so far. 2 AS number corrections 1 non-operational connection listed as operational 6 operational connection not listed (4 involving CIX/PAIX) So far the most common response has concerned the CIX =/= PAIX problem. When I set up my spreadsheet, there was only the 149.20 CIX network, and that's what the spreadsheet is based on. The CIX router at the PAIX is also connected to the PAIX FDDI ring. After some mail last night, I looked at adding the PAIX side; but it quickly got very complicated. There are some people on the PAIX side which are peered with the CIX router, and some that aren't. But I can't tell just by pinging the network address. Then I started getting mail asking why I hadn't included a variety of the other 200+ exchange points that claim to be operating world-wide. So the question becomes, if I do this again, how do I choose the six (because I have limited resources, and no government grant money, and can't get advertising dollars publishing an ISP directory) major, biggest, most important,or just plain interesting exchange points in the world to put the the spreadsheet. Or is the entire thing irrelevant, because everyone is moving to private bilateral connections. And the important thing is how many providers does someone peer, not how many exchange points they're connected to. For the price of one Sprint-NAP connection, I can get several connections to Canada. And we have a lot more customers in Canada than at the Sprint-NAP. We already peer with everyone at the Sprint-NAP, or been turned down by them elsewhere. So one more exchange point doesn't buy much. -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

3 2

Test Message
by jmbrown＠ihighway.net 05 Apr '97

05 Apr '97

This is a test message. Please IGNORE

1 0