- NANOG - lists.nanog.org

Re (2): Core router bakeoff?
by jkucic＠kpmg.com 09 May '98

09 May '98

Eric Germann <ekgermann(a)cctec.com>,Internet writes: Given that Cisco built their routers for them, and then they pissed off Cisco to the point where they didn't renew there code license for IOS in there hubs because of their anti Cisco marketing campaign. It was a limited 4000 Router. Only two slots instead of three. Cabletron is now pushing their new SmartSwitch Router that they acquired when they purchased YAGO systems. They claim its an ISP backbone quality router (honest its on their web page). But I don't believe it does BGP4. Cabletron has a bad rap with consultants. Don't ever try to partner with them and walk them in someplace. They are the borg, you will be assimilated. Hey, you haven't spoken with your local Cabletron rep lately; Cabletron is pushing channel relationships again. Now that I ragged on em, they do have solid stuff. I have one client who has had their MMAC hubs up and running for about 8 years with narry a hiccup. And these are in rotten basements. You are right but is it Year 2000 compliant? Thanks, Joe Kucic

5 4

Re: Core router bakeoff?
by shields＠crosslink.net 09 May '98

09 May '98

In article <199805081728.NAA19463(a)persephone.ny.ans.net>, Selina Priestley <selina(a)ans.net> wrote: > One PC remote management win is that you can actually have secure > encrypted access/authentication for remote management, rather than > relying passwords in the clear and tftp. This is lacking in all > router venders at this instant, far as I know. Cisco has support for Kerberized telnet in some images -- good luck getting it to work, though. -- Shields, CrossLink.

1 0

Re: Core router bakeoff?
by Selina Priestley 09 May '98

09 May '98

> > In short, unless you have numbers to indicate otherwise, I can't say > that PCs cost you much, in the short term or the long term. > > What Cisco's will buy you is better handling of T3s, SONET, etc., and > the ability to handle much higher performance lines. They also have > much better remote management facilities, in so far as the things have > real serial consoles and such, which PCs don't. > > Perry One PC remote management win is that you can actually have secure encrypted access/authentication for remote management, rather than relying passwords in the clear and tftp. This is lacking in all router venders at this instant, far as I know. Selina

3 2

RE: Microsoft NOC?
by Kurt Parent 09 May '98

09 May '98

Microsoft's NOC is best accessed via the msnnoc(a)microsoft.com e-mail alias. The peering(a)microsoft.com alias handles (what else) peering inquiries and internal communication. thanks, kurt- Kurt Parent kparent(a)microsoft.com > -----Original Message----- > From: Reid B. Fishler [SMTP:redhead@cloudy.lightning.net] > Sent: Thursday, May 07, 1998 3:44 PM > To: kit(a)connectnet.com > Cc: nanog(a)merit.edu > Subject: Re: Microsoft NOC? > > try writing peering(a)microsoft.com. I got a live person there...In their > networking dept. > > Reid > > > > > > 425-882-8080 as listed by whois for Microsoft is not their NOC. When > you > > call this number you are told that they will only transfer by a persons > > name or extension and offer no other assitance. > > > > If anyone has any other contact information it would be appreciated. > > > > -- > > > > Kit Knox [kit(a)connectnet.com] - CONNECTnet INS, Inc. > >

6 5

sprint's outage list
by Dean Gaudet 09 May '98

09 May '98

Ya know, I really hate to do this. But I've tried through all the correct/expected channels. But sprint won't remove me from their outage list. I've mailed majordomo(a)sprint.net @sprintlink.net @mercury.res.sprintlink.net, and they all claim I'm not a list member. I've mailed postmaster, owner-outage, outage-owner, postout, outage-request. I even sent it to outage-outgoing, even though I knew it wouldn't get through. I've sent the request to the tech contact for sprint.net, which is the same as the tech contact for the reverse DNS of a few of their nets, and is the same as the address listed in their SOA records. I sent it to the tech contact for AS1239 (noc(a)sprint.net) I considered sending it to all the mail contacts for MAINT-AS1239 (radb), but most of them don't even appear to be sprint employees. As you can see from the headers below, the message is delivered to dgaudet-list-outage(a)arctic.org. I've tried all of the above using that as the envelope sender -- I'm not just some stupid luser that doesn't understand this crap. All to no avail. So here I come, hoping to embarass some sprint employee to take me off their damn list. I'm sorry to waste the rest of your time. Thanks Dean ---------- Forwarded message ---------- Return-Path: <owner-outage(a)sprint.net> Delivered-To: dgaudet-list-outage(a)arctic.org Received: (qmail 15150 invoked from network); 9 May 1998 08:07:29 -0000 Received: from gate1.sprintlink.net (199.0.233.2) by twinlark.arctic.org with SMTP; 9 May 1998 08:07:29 -0000 Received: from mercury.res.sprintlink.net by gate1.sprintlink.net via smtpd (for twinlark.arctic.org [204.62.130.91]) with SMTP; 9 May 1998 08:07:29 UT Received: (from root@localhost) by mercury.res.sprintlink.net (8.7.3/8.6.12) id DAA05866 for outage-outgoing; Sat, 9 May 1998 03:24:57 -0400 (EDT) Received: from elle.res.sprintlink.net (elle.res.sprintlink.net [199.0.234.34]) by mercury.res.sprintlink.net (8.7.3/8.6.12) with ESMTP id DAA05861 for <outage(a)mercury.res.sprintlink.net>; Sat, 9 May 1998 03:24:56 -0400 (EDT) Received: from gate1dmz.res.sprintlink.net (gate1dmz.res.sprintlink.net [199.0.234.1]) by elle.res.sprintlink.net (8.8.5/8.8.5) with SMTP id DAA17852 for <outage(a)sprint.net>; Sat, 9 May 1998 03:51:23 -0400 (EDT) Received: from jay.sprintlink.net ([199.0.233.8]) by gate1dmz.res.sprintlink.net via smtpd (for elle.res.sprintlink.net [199.0.234.34]) with SMTP; 9 May 1998 07:49:53 UT Received: from gate1.sprintlink.net (gate1.sprintlink.net [199.0.233.2]) by kay.sprintlink.net (8.8.5/0.0.0) with SMTP id DAA21247 for <outage(a)sprint.net>; Sat, 9 May 1998 03:50:27 -0400 (EDT) Received: from vas.res.sprintlink.net by gate1.sprintlink.net via smtpd (for jay.sprintlink.net [199.0.233.8]) with SMTP; 9 May 1998 07:49:52 UT Received: (from httpd@localhost) by vas.res.sprintlink.net (8.8.5/8.6.12) id DAA13839; Sat, 9 May 1998 03:48:49 -0400 (EDT) Date: Sat, 9 May 1998 03:48:49 -0400 (EDT) Message-Id: <199805090748.DAA13839(a)vas.res.sprintlink.net> From: Outage Poster <postout(a)sprint.net> To: outage(a)sprint.net Subject: Outage in Fort Worth area Sender: owner-outage(a)sprint.net Precedence: bulk Sat May 9 03:48:48 EDT 1998 OUTAGE MESSAGE The problem that was reported of the outages in the Fort Worth area (@ 00:05 EDT on 5/09/98) has been fixed. There should not be any more problems in the area. Thank you for bearing with us. Andrew Jacobsen, -Sprintlink NOC ===================================================================== For Sprintlink network status, call 1-800-232-6895; option 1, 1. For Sprintlink scheduled maintenance activities, go to http://www.sprint.net/maint/

1 0

Re: Core router bakeoff?
by Tony Li 09 May '98

09 May '98

perry(a)piermont.com (Perry E. Metzger) writes: > > Interesting. Have you done any scalability testing? Per-flow state has > > been shown to scale poorly in Internet backbones. > > I agree with you that it doesn't scale well to huge routers. I am > certain that it would work very badly in MCI's backbone. It probably > will work just fine in a small provider, though, and those are most of > the guys who want to roll their own out of PCs anyway. Not a foregone conclusion. Recall that it's not the absolute volume of the traffic that's the difficulty but the diversity and 'hit ratio' of the particular traffic patterns. If the flow cache is small and stable, then you could push terabits. The problem that has been seen is that anything connected to the Internet with a significant community behind it has critical mass to churn the flow cache. > If you want to pump half a gigabit through on a giant peering router, > caches will lose, and the only thing that will do you right is > real router hardware. I wouldn't know anything about that. Tony

1 0

Re: Core router bakeoff?
by Tony Li 09 May '98

09 May '98

paul(a)vix.com (Paul Vixie) writes: > > > BTW, NetBSD with the recent flow cache mods can handle at least > > > 150,000 packets per second. We haven't seen what the actual upper > > > limit is, but that number doesn't seem to be eating a lot of CPU. > > > > Interesting. Have you done any scalability testing? Per-flow state has > > been shown to scale poorly in Internet backbones. > > That was with one flow. The idea of "there basically aren't flows" is not > obvious on first approach to these problems, wouldn't you say, Tony :-)? Not obvious to some, but we sat and watched a significant router vendor make that mistake (despite good advice to the contrary) anyhow. I'd hate to see someone else repeat that mistake. Tony

1 0

Re: Core router bakeoff?
by Blake Willis 09 May '98

09 May '98

Jay, You're right. You win, to a certain degree. To quote Livingston online docs: ComOS release 3.5 now supports Variable Length Subnet Masks. In previous releases ComOS required the same netmask to be used for all subnets of a network. <snip snip> I believe that this was also the first ComOS to support OSPF. Have fun with your router... -Blake --------------------------------------------------------------------------- Blake Willis 703-448-4470x483 Network Engineer, New Customers blakew(a)cais.net CAIS Internet, a CGX Communications Company --------------------------------------------------------------------------- On Fri, 8 May 1998, Jay 'Whip' Grizzard wrote: > > Command> ver > > Livingston PortMaster IRX ComOS 3.5R > > System uptime is 269 days 20 hours 2 minutes > > man-nh-gw2> ver > Livingston Enterprises PortMaster Version 3.2.1R > System uptime is 708 days 3 hours 56 minutes > > ... I win. :>

2 1

Re: Re (2): Core router bakeoff?
by jkucic＠kpmg.com 09 May '98

09 May '98

Hi, Eric Germann <ekgermann(a)CCTEC.COM>,Internet writes: >Do we really care if an unmanaged hub is Y2K compliant? As long as it >forwards packets.... I don't but its an opportunity to sell the client. And if you don't test it I recommend you don't get quoted saying it doesn't matter unless you want a lawsuit if anything happens. Customers want everything in writing to be Year 200 compliant or not having any affect on their operation and in many cases regulators are mandating it in writing. Thanks, Joe Kucic

3 2

NOC Phone Number Time (Re: Microsoft NOC?)
by Sean Donelan 09 May '98

09 May '98

>The reason it is imperative that there be a telephone number for an >ISP is that it's really hard to use e-mail to contact an ISP if the >network connectivity to the ISP is not working. I guess its that time again. Between May 11 and May 15 I will be conducting my Emergency NOC Phone Test. This is an unannounced test where I call each of the phone numbers given to me by providers I interconnect with. The rules are listed on <http://dranet.dra.com/verify-contact.html>. Although I say its 'unannounced' if you want to tell your NOC folks I'll be calling next week go ahead. In practice I've found corporate culture tends to overwhelm any special advance notice. In case you are interested, the very best performance in the last emergency NOC phone contact test was by the Defense Communications Agency/DISA NOC. They earned four stars, out of five, losing only one point because the operator did not know how to call their NOC from outside the US. 1-800 numbers are difficult to dial outside the US. -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Affiliation given for identification not representation

1 0