- NANOG - lists.nanog.org

BGP evaluation
by Sathya Perla 09 Aug '02

09 Aug '02

Hi, In general, how is a BGP implementation evaluated/benchmarked by an operator? Are there any well-known figures relating to memory-consumption, speed of convergence, scalability etc. that one can benchmark with? There was a write-up on Data Comm magazine testing Juniper's M40 BGP. I haven't seen much more than that - especially test procedures used to compare various vendor implementations. Thanks in advance, -Sathya

2 1

Re: NSPs filter?
by Stephen Stuart 09 Aug '02

09 Aug '02

> One thing that sometimes comes up is that people do number links using > RFC1918 address space which occasionally results in an ICMP 'fragmentation > needed but DF bit set' packet with an RFC1918 source address. Filtering out > this packet could result in TCP breaking. That can be accomodated; behold, all the joy of PMTUD, with none of the other crap from designated special-use address space: firewall { family inet { filter external-filter { term allow-icmp-unreach { from { protocol icmp; icmp-type unreachable; icmp-code fragmentation-needed; } then { count allow-icmp-need-frag; accept; } } term allow-icmp-timxceed { from { protocol icmp; icmp-type time-exceeded; icmp-code [ ttl-eq-zero-during-transit ttl-eq-zero-during-reassembly ]; } then { count allow-icmp-timxceed; accept; } } term deny-rfc1918 { from { source-address { 10.0.0.0/8; 172.16.0.0/12; 192.168.0.0/16; } } then { count deny-rfc1918; discard; } } term deny-test { from { source-address { 192.0.2.0/24; } } then { count deny-test-net; discard; } } term deny-autoconfig { from { source-address { 169.254.0.0/16; } } then { count deny-autoconfig; discard; } } term LAST { then accept; } } } } Application is left as an exercise to the reader. Stephen

1 0

Re: Re: NSPs filter?
by sal_sabella＠hushmail.com 09 Aug '02

09 Aug '02

Stephen Stuart wrote: >I made what I consider to be non-trivial progress toward stamping that >out before my time at the helm of 6461 ended. I happen to know that >the people still there are trying to continue making progress on that >front, what is going on around them. Confucious say: Man who sit on rose bush soon get point. >Would you care to take a shot at answering my question, or is >contributing productively too much to ask? My employer believes against filtering on source or destination. Communicate in total privacy. Get your free encrypted email at https://www.hushmail.com/?l=2 Looking for a good deal on a domain name? http://www.hush.com/partners/offers.cgi?id=domainpeople

4 3

Re: your mail
by Ukyo Kuonji 09 Aug '02

09 Aug '02

Sorry, but there are valid reasons for posting anonymously. As one person that almost got fired from a company for expressing what was clearly marked as his own opinion to this list, I can attest to this. Some employers are concerned that you post during working hours, some are concerned that you use a work address, some are concerned that anything you say, no matter what email you use, will be considered as representations of the company's direction, even if it is clearly marked as not. I like helping and sharing my opinions, but I'm not willing to loose my job over it. UK >Date: Thu, 8 Aug 2002 00:13:22 -0400 (EDT) >From: Brian Wallingford <brian(a)meganet.net> > >Perhaps it's time to bar posting privileges from those who insist on >remaining entirely anonymous? > >I doubt that anyone who has anything substantive to offer would need to >use a hushmail/yahoo/etc. return address. > >The initial posts from Bandy, Vaul, et al were mildly amusing at first, >but the novelty wore out very quickly. _________________________________________________________________ Send and receive Hotmail on your mobile device: http://mobile.msn.com

3 2

[lamour@mail.argfrp.us.uu.net: Fwd: Re: If you have nothing to hide]
by Todd MacDermid 08 Aug '02

08 Aug '02

In message <20020805225221.82473.qmail(a)sidehack.sat.gweep.net>, bdragon(a)gweep.n et writes: > >I was not aware that responses to source-routed packets were themselves >source-routed. I also don't believe it is the case, but am open to being >contradicted. If the responses aren't source-routed, then the packets would >only return through your network if your network was the path back to the >spoofed source. A friend of mine directed me to this thread. Source routed packets can indeed be used to spoof IP connections, and I've written a tool to do it. It's available at http://www.synacklabs.net/projects/lsrtunnel If you simply want to check host behaviour to see if you can spoof connections, I've written a scanner at http://www.synacklabs.net/projects/lsrscan Short story is Solaris < 8 will reverse source routes by default, and Windows boxes will reverse source routes by default. The BSDs and Linuces I've tested mostly block source routed packets by default. Todd

1 0

Anyone heard of anything at OC24?
by Ryan Mooney 08 Aug '02

08 Aug '02

Hi, Hopefully this isn't to far off topic, but I have sort of a wierd problem that no one else has been able to figure out and was wondering if the collective intellect gathered here might be able to drum up a solution. I work for a company that has a wireless product that we've been developing that is basically a fiber repeater. The limiting factors are our transmitter and detector circuitry capabilities and the clock recovery circuitry. Right now we can clock the units to OC12, OC24 and GE (1.25Gbps). I have all of the management types wanting to demonstrate "gigabit atm" or "gigabit pos", and we won't be able to do OC48 for a while. So at this point the closest we can get is OC24 (not counting GE, since we have some customers that really want ATM for legacy reasons). While we can do OC12 they aren't happy w/ that since its less than 1Gbps in one direction (and well, whatever I'm sure you know how irrational discusions like that go). The problem I'm having is that I can't find anything that does anything at OC24. If I could find a simple TDM device or circuit that combined 2xOC12 to OC24 that would be perfect (and stays in line with the protocol agnostic radio theory). Failing that some sort of ATM device that has OCx and OC24 interfaces would be a nifty second choice. If any of you know of any chipsets and/or products that do this let me know. Please send replies direct to me as I would be suprised if the answers are of much interest to anyone else (if I'm wrong on that one, let me know and I'll send you a summarization). Thanks. -- >-=-=-=-=-=-=-<>-=-=-=-=-=-<>-=-=-=-=-=-<>-=-=-=-=-=-<>-=-=-=-=-=-=-< Ryan Mooney ryan(a)pcslink.com <-=-=-=-=-=-=-><-=-=-=-=-=-><-=-=-=-=-=-><-=-=-=-=-=-><-=-=-=-=-=-=->

1 0

NANOG, its decline in s/n
by John M. Brown 08 Aug '02

08 Aug '02

Having been a participant of the NANOG list since 1995, twice a past host of the meetings, and someone that believes we should help and educate, I am most concerned over the trend that is showing itself here on the list. This is a list about knowledge, share, help, and operations. Some on this list are "Old Pro's" others are just "Old ....." I see new names, new faces, sometimes the same questions, sometimes slight variences of those questions. What then happens is people attack the sender (heck I've even been slightly guilty of this), banish them to newbie.dev.nul. Threads move quickly from the question to debates on random junk and thus become a waste to the community. When I first got on the Internet, 1984, it was full of cooperative and helpful people. 18 years later, not so much. We have real issues to deal with, DMCA, Clarke, The Implosion of ICANN (with means the ITU will run things me thinks, not good), and various other laws, acts and general clue.void from .void.gov. So instead of picking on each other, calling it "standard NANOG hazing" or whatever, lets help, be good in our technical honesty, or don't post. If someone doesn't listen to your answer, a BIGGER stick isn't going to help, take it private, or move along. We are all pro's here, right ?? We got a net to run, lets go do it. respectfully, john brown a person not speaking for anyone else

9 13

well the rt's have it:)
by Scott Granados 08 Aug '02

08 Aug '02

By far it seemed most were using rt and remedy as #2. Thank you for everyone who responded I'd thank everybody individually as I probably should:) but I got over 100 responses so here's a blanket thanks:)

1 0

BGP, How-tos, tips and tricks, training info.
by crimson＠unspeakable.org 08 Aug '02

08 Aug '02

Does anyone have a good url/documentation/howto's about BGP and implementing it. I'm looking for how gto implement it, considersations, examples, common use commands, troubleshooting. Basically I want to get up to speed on BGP as my company is looking into rolling it out. And I don't want to waste everyone's time dealing with yet another site on the net with messed up BGP :) Please reply off list as to cut down on the crap posted on it. So no one gets there panties in a knot.

4 3

Re: Do ATM-based Exchange Points make sense anymore?
by michael.dillon＠radianz.com 08 Aug '02

08 Aug '02

Given that your analysis could be used to make decisions about exchange point architecture, I wonder if you would consider adding RPR IEEE 802.17 into the mix? While I don't know of anyone using RPR in an exchange point, there certainly are products on the market and not just Cisco's SRP/DPT. Within Ebone we were using SRP/DPT rings quite successfully in our PoP architecture, i.e. it does work in the real world. I know that the marketing side of RPR is promoting it as a metro area solution, but it works just as well in a PoP. Given that you are comparing a PoP oriented technology (Ethernet) with a metro area technology (ATM), I think it makes sense to take a serious look at a technology that covers both. --Michael Dillon

4 3