[routing-wg]Idea about AS-usage on IXPs
Hi, Since all public IXPs are using unique IP-Allocations for the exchange fabric, I thought about the possibility to hook them all up under the same AS-number. Each prefix/IXP is visible the same as before and exchange networks could be easily identified by the AS-number and specially treated (filtered/preferred). Feedback? Thanks, Kurt
On 01.12.2006 13:18 Kurt Kayser wrote
Hi,
Since all public IXPs are using unique IP-Allocations for the exchange fabric, I thought about the possibility to hook them all up under the same AS-number.
Each prefix/IXP is visible the same as before and exchange networks could be easily identified by the AS-number and specially treated (filtered/preferred).
Feedback?
And how do I reach AMS-IX or LINX (or any other IXP)? Arnold -- Arnold Nipper / DE-CIX Management GmbH, the German Internet Exchange e-mail: arnold.nipper@de-cix.net phone/mob: +49 172 265 0958 fax: +49 6224 9259 333 http://www.de-cix.net pgp-fingerprint: 393E 1B14 AE49 816F 7EA3 87F6 AF3E 82C2 9441 FC4C
Hi Arnold, Arnold Nipper wrote:
And how do I reach AMS-IX or LINX (or any other IXP)?
Is inter-IXP-traffic really typical/necessary? Active devices with public content could be connected to general ISP-Networks. Service-machines are mostly anyway for connected parties of particular interest. Regards, Kurt
Is inter-IXP-traffic really typical/necessary?
from a router on IX A, i want a nice clean traceroute to a router on IX B, and only have access to that IX via my friend's network on IX A. randy
Hi Randy, Randy Bush wrote:
Is inter-IXP-traffic really typical/necessary?
from a router on IX A, i want a nice clean traceroute to a router on IX B, and only have access to that IX via my friend's network on IX A.
If the router is not an IXP-owned router, each participating router is in a "normal" AS. But I might be really wrong in order to save a couple of 2-Byte AS-numbers, but somehow I have the feeling that 2-Byte space is out much faster that anticipated. So how far are we in globally reachable 4-Byte AS-numbers yet? (different topic). Regards, Kurt
Hi, On Fri, Dec 01, 2006 at 02:18:40PM +0100, Kurt Kayser wrote:
But I might be really wrong in order to save a couple of 2-Byte AS-numbers, but somehow I have the feeling that 2-Byte space is out much faster that anticipated.
Given the number of (recognized) IXes, the number of 2-Byte AS numbers, and the number of "lost" AS numbers, I'd say that focussing on IXes is not going to gain very much - but possibly causing lots of confusion and problems in troubleshooting. Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 98999 SpaceNet AG Mail: netmaster@Space.Net Joseph-Dollinger-Bogen 14 Tel : +49-89-32356-0 D- 80807 Muenchen Fax : +49-89-32356-234
At 12:18 AM 2/12/2006, Kurt Kayser wrote:
Hi Randy,
Randy Bush wrote:
Is inter-IXP-traffic really typical/necessary?
from a router on IX A, i want a nice clean traceroute to a router on IX B, and only have access to that IX via my friend's network on IX A.
If the router is not an IXP-owned router, each participating router is in a "normal" AS.
But I might be really wrong in order to save a couple of 2-Byte AS-numbers, but somehow I have the feeling that 2-Byte space is out much faster that anticipated.
Nope - the 2byte number consumption is pretty much right on schedule.
So how far are we in globally reachable 4-Byte AS-numbers yet? (different topic).
Ask your vendor about the state of their code! Geoff
On Fri, Dec 01, 2006 at 02:18:40PM +0100, Kurt Kayser wrote:
Hi Randy,
Randy Bush wrote:
Is inter-IXP-traffic really typical/necessary?
from a router on IX A, i want a nice clean traceroute to a router on IX B, and only have access to that IX via my friend's network on IX A.
If the router is not an IXP-owned router, each participating router is in a "normal" AS.
But I might be really wrong in order to save a couple of 2-Byte AS-numbers, but somehow I have the feeling that 2-Byte space is out much faster that anticipated. So how far are we in globally reachable 4-Byte AS-numbers yet? (different topic).
According to RIPE-389 (http://www.ripe.net/docs/asn-assignment.html#19): 1.9 4-Byte AS Numbers RIPE NCC will assign 4-Byte AS Numbers according to the following timeline: * From 1 January 2007 the RIPE NCC will process applications that specifically request 4-byte only AS Numbers and assign such AS Numbers as requested by the applicant. In the absence of any specific request for a 4-byte only AS Number, a 2-byte only AS Number will be assigned by the RIPE NCC. * From 1 January 2009 the RIPE NCC will process applications that specifically request 2-byte only AS Numbers and assign such AS Numbers as requested by the applicant. In the absence of any specific request for a 2-byte only AS Number, a 4-byte only AS Number will be assigned by the RIPE NCC. * From 1 January 2010 the RIPE NCC will cease to make any distinction between 2-byte only AS Numbers and 4-byte only AS Numbers, and will operate AS Number assignments from an undifferentiated 4- byte AS Number allocation pool. Terminology "2-byte only AS Numbers" refers to AS Numbers in the range 0 - 65535 "4-byte only AS Numbers" refers to AS Numbers in the range 1.0 - 65535.65535 (decimal range 65,536 - 4,294,967,295) "4-byte AS Numbers" refers to AS Numbers in the range 0.0 - 65535.65535 (decimal range 0 - 4,294,967,295) -- Regards, Volodymyr.
Hi! By the way, is 0.11111 and 11111 same ASes? Volodymyr Yakovenko wrote:
On Fri, Dec 01, 2006 at 02:18:40PM +0100, Kurt Kayser wrote:
Hi Randy,
Randy Bush wrote:
Is inter-IXP-traffic really typical/necessary? from a router on IX A, i want a nice clean traceroute to a router on IX B, and only have access to that IX via my friend's network on IX A. If the router is not an IXP-owned router, each participating router is in a "normal" AS.
But I might be really wrong in order to save a couple of 2-Byte AS-numbers, but somehow I have the feeling that 2-Byte space is out much faster that anticipated. So how far are we in globally reachable 4-Byte AS-numbers yet? (different topic).
According to RIPE-389 (http://www.ripe.net/docs/asn-assignment.html#19):
1.9 4-Byte AS Numbers
RIPE NCC will assign 4-Byte AS Numbers according to the following timeline:
* From 1 January 2007 the RIPE NCC will process applications that specifically request 4-byte only AS Numbers and assign such AS Numbers as requested by the applicant. In the absence of any specific request for a 4-byte only AS Number, a 2-byte only AS Number will be assigned by the RIPE NCC. * From 1 January 2009 the RIPE NCC will process applications that specifically request 2-byte only AS Numbers and assign such AS Numbers as requested by the applicant. In the absence of any specific request for a 2-byte only AS Number, a 4-byte only AS Number will be assigned by the RIPE NCC. * From 1 January 2010 the RIPE NCC will cease to make any distinction between 2-byte only AS Numbers and 4-byte only AS Numbers, and will operate AS Number assignments from an undifferentiated 4- byte AS Number allocation pool.
Terminology
"2-byte only AS Numbers" refers to AS Numbers in the range 0 - 65535
"4-byte only AS Numbers" refers to AS Numbers in the range 1.0 - 65535.65535 (decimal range 65,536 - 4,294,967,295)
"4-byte AS Numbers" refers to AS Numbers in the range 0.0 - 65535.65535 (decimal range 0 - 4,294,967,295)
-- WBR, Max Tulyev (MT6561-RIPE, 2:463/253@FIDO)
Hi, Kurt Kayser schrieb:
Hi,
Since all public IXPs are using unique IP-Allocations for the exchange fabric, I thought about the possibility to hook them all up under the same AS-number.
Each prefix/IXP is visible the same as before and exchange networks could be easily identified by the AS-number and specially treated (filtered/preferred).
Feedback?
i don't see any point in that suggestion... Not all IXes will do that anyways (how to force them?), saving AS numbers is not an issue (anymore), so it's a PITA at best when it comes to debugging some issues with a certain IX. Any why would anyone treat/filter/prefer ALL Exchanges around the world the same way? ...but probably i just miss the point, the suggestion is not elaborate enough :-) -- ======================================================================== = Sascha Lenz SLZ-RIPE slz@baycix.de = = Network Operations = = BayCIX GmbH, Landshut * PGP public Key on demand * = ========================================================================
Kurt, which problem do you try to solve? :-) In principle an IXP is a Layer 2 fabric with no IP stuff involved. If there is Layer 3 stuff (route servers, statistics, web,...) then the *IP* stuff would usually have an IXP-defined routing policy. Why should we try to force indepentend organisations to use the same routing policy all over the places? Cheers, servus, Wifried. Kurt Kayser wrote:
Hi,
Since all public IXPs are using unique IP-Allocations for the exchange fabric, I thought about the possibility to hook them all up under the same AS-number.
Each prefix/IXP is visible the same as before and exchange networks could be easily identified by the AS-number and specially treated (filtered/preferred).
Feedback?
Thanks, Kurt
participants (9)
-
Arnold Nipper -
Geoff Huston -
Gert Doering -
Kurt Kayser -
Max Tulyev -
Randy Bush -
Sascha Lenz -
Volodymyr Yakovenko -
Wilfried Woeber, UniVie/ACOnet