I apologize if this question is not relevant to your discussions, but Susan Harris, e-mail address srh@merit.edu, suggested that I ask it here... The peering table that appears on the London Internet Exchange web site www.linx.net/peering.html...... is this something that can be generated for different peering locations? Also with this information would it be possible to create a table that lists companies and the cities that they are peering in? Any insight would be of great help to me thank you. Once again any advice on how or where to obtain this info would be of great benefit to me. Thanks Jeff Lentz
On Wed, 19 Apr 2000, Jeff Lentz wrote:
I apologize if this question is not relevant to your discussions, but Susan Harris, e-mail address srh@merit.edu, suggested that I ask it here...
The peering table that appears on the London Internet Exchange web site www.linx.net/peering.html...... is this something that can be generated for different peering locations? Also with this information would it be possible to create a table that lists companies and the cities that they are peering in? Any insight would be of great help to me thank you.
The peering matrix at LINX doesn't necessarily mean that any parties shown to peer actually do it across the LINX exchange point fabric itself, they could be PIs anywhere in the world or peerings at other IXPs. Why is this? Well, take a read of the footnotes: "This table is generated from objects in the RIPE and RA databases. If there is something missing or wrong then please refer to the original data in the databases. The information given here can apply to peering arrangements between members anywhere in the world, and is not limited to the traffic flowing through the LINX." It's basically a script which pulls all the necessary aut-num objects from RIPE and RADB and builds a table out of the as-in/as-out declarations. It's just taking information which is available in the public domain (if you know how to drive a whois client) and munging it into a tabluar form. So, in response to JSB's comment about confidentiality of peering arrangements: Yep, sure, you can be confidential about where you interconnect, at what speed and choice of media, should you choose. But, you can't be confidential about who you are peering with, because the RA/RIPE will tell anyone with a bit of clue. Unless, of course, you aren't declaring these "top secret" peerings in the relevant routing database(s), which is probably a bad thing [tm], but that's my personal view on it. Also, some locations, such as LINX and PAIX, do have publically accessible member info on their websites, so you get a good idea of who is connected at that exchange point. There are, as JSB suggests, plenty of tools out there to help you with routing/peering questions. This is me with my own hat on, BTW, not my LINX engineer hat. Mike
"This table is generated from objects in the RIPE and RA databases. If there is something missing or wrong then please refer to the original data in the databases.
The information given here can apply to peering arrangements between members anywhere in the world, and is not limited to the traffic flowing through the LINX."
When the peering matrix perl script was written (false modesty be damned :), we decided that it was too complicated to bother with anything other than the public information in the RA/RIPE db's. The other though was that if two AS's peer someone else and not at the LINX, then that is their loss, and not one that we should have tried too hard to fix for them. The use of the RA/RIPE db's was seen as the only "neutral" data available, and neutrality was (and I think still is) a priority at the LINX. Peter
On 19 Apr 00, at 17:04, Peter Galbavy wrote:
When the peering matrix perl script was written (false modesty be damned :),
I would normally give credit where credit is due - but I wasn't 100% sure if you authored said piece of perl (but had a good idea), and didn't want to offend :-).
The use of the RA/RIPE db's was seen as the only "neutral" data available, and neutrality was (and I think still is) a priority at the LINX.
Yes, it most certainly is! Mike -- Mike Hughes,London,UK http://www.smashing.net/ mike@smashing.net ICQ UIN:4871914 "Only one thing in life is certain: init is process No. 1" PGP Public Keys from http://www.smashing.net/pgpkeys.html
On Wed, Apr 19, 2000 at 04:48:14PM +0100, Mike Hughes wrote:
Yep, sure, you can be confidential about where you interconnect, at what speed and choice of media, should you choose. But, you can't be confidential about who you are peering with, because the RA/RIPE will tell anyone with a bit of clue.
Only if people register correct information. Even with the coming implementation of the RPSL Security RFC, nothing can be done about representations for import and export policies in the aut-num objects. The best you can do is note inconsistant policy and trust things a bit more if the policy is consistant. A better indication comes from browsing the AS Paths available in the global BGP. While it is technically possible (shudder) to forge information in the AS Path, this is the best indication of where two AS's have an adjacency. Of course, it doesn't say anything about where that adjacency is. FWIW, RPSL makes it easier than RIPE-181 to specify exceptions to your global policy at particular interfaces. However, most people that register in the public IRR do so to disclose global policy rather than per-router policy. Its nice to see that various parties are documenting operationally useful things in the IRR, such as community strings. I would be slighly surprised if these types of AS regex's are actually used in practice (not to pick on Exodus): import: from AS3967 accept <^AS-EXODUS*$> AND NOT {0.0.0.0/0} Regex matching can be quite processor intensive, and as-exodus isn't small.
Mike
-- Jeffrey Haas - Merit RSng project - jeffhaas@merit.edu
I apologize if this question is not relevant to your discussions, but Susan Harris, e-mail address srh@merit.edu, suggested that I ask it here...
The peering table that appears on the London Internet Exchange web site www.linx.net/peering.html...... is this something that can be generated for different peering locations? Also with this information would it be possible to create a table that lists companies and the cities that they are peering in? Any insight would be of great help to me thank you.
Once again any advice on how or where to obtain this info would be of great benefit to me. Thanks
Jeff Lentz
In the few places I've participated in, this type of data is often considered confidential between parties. In some locations, like LINX, publication of this type of data is considered as part of the service. It really depends on the exchange operator and the participants at that exchange. That being said, there are a number of good efforts to debate discuss and publish data on this topic. Ren (Onyx) and WBN (Equinix) are fairly active in this arena. In fact, I think WBN will be hosting yet another peering BOF/pannel at the next NANOG mtg.
participants (5)
-
bmanning＠vacation.karoshi.com -
Jeff Haas -
Jeff Lentz -
Mike Hughes -
Peter Galbavy