I've noticed that http://www.nanog.org/filter.html is a little out of date, and I was wondering if anyone knows whether any ISPs currnetly have issues with /24 BGP advertisements for blocks outside of the traditional Class C space? We have some migrations to do from one space to another and having the ability to do some /24 advertisements during that period would be greatly helpful. I see that even Verio now accepts prefix lengths /24 or shorter. Thanks, -- Adam Clark Senior Network Engineer Canada Web Hosting
Adam Clark said the following on 24/8/07 04:02:
I've noticed that http://www.nanog.org/filter.html is a little out of date, and I was wondering if anyone knows whether any ISPs currnetly have issues with /24 BGP advertisements for blocks outside of the traditional Class C space?
My regular analysis report shows 120790 /24s in the BGP table today. So I'd suspect that there are few problems with /24 advertisements; but that's just my BGP view, could be different elsewhere. Before you start your migration, announce one /24 as a trial and see what happens? philip --
On Fri, Aug 24, 2007 at 06:14:27AM +1000, Philip Smith wrote:
Adam Clark said the following on 24/8/07 04:02:
I've noticed that http://www.nanog.org/filter.html is a little out of date, and I was wondering if anyone knows whether any ISPs currnetly have issues with /24 BGP advertisements for blocks outside of the traditional Class C space?
My regular analysis report shows 120790 /24s in the BGP table today. So I'd suspect that there are few problems with /24 advertisements; but that's just my BGP view, could be different elsewhere.
I've got an interesting view/project that i've been working on that is related to announcements ... ;) you need to watch what you announce even if it's transient, you may end up on my webpage.. i've only got about a day or two of data, but if you check out my page here: http://puck.nether.net/bgp/leakinfo.cgi?search=do&src=nanog You'll see recent routing leaks as well as be able to search them. It's still in "alpha" mode, but you may find this interesting to search on your as-path to find out who may be leaking your routes. Some folks have complained that I don't currently exempt a couple of the RIPE RIS beacon prefixes, so if you see those, it may not be an issue. btw, many thanks to those networks who have already cleaned up or responded to questions about your policy in the pre-alpha stage of this project. - Jared -- Jared Mauch | pgp key available via finger from jared@puck.nether.net clue++; | http://puck.nether.net/~jared/ My statements are only mine.
On Thu, Aug 23, 2007 at 11:02:28AM -0700, Adam Clark wrote: [snip]
We have some migrations to do from one space to another and having the ability to do some /24 advertisements during that period would be greatly helpful.
Always assume you have no visibility everywhere and that your squeakiest wheel will have some connection to a site you hadn't considered, believed obscure, etc. Recall the first clause of the robustness principle and apply it to your signalling. Test well in advance of any actual cut. -- RSUC / GweepNet / Spunk / FnB / Usenix / SAGE
We have some migrations to do from one space to another and having the ability to do some /24 advertisements during that period would be greatly helpful. Always assume you have no visibility everywhere and that your squeakiest wheel will have some connection to a site you hadn't considered, believed obscure, etc. Recall the first clause of the robustness principle and apply it to your signalling. Test well in advance of any actual cut.
how? if i read you aright, you are saying that there will likely be a few strange folk at the 'edges' of the internet who will have problems and whine. but it is extremely hard to get data (i.e. route views, ris, traceroute servers, ...) from the edge. randy
On Thu, 23 Aug 2007 11:27:31 -1000, Randy Bush said:
how? if i read you aright, you are saying that there will likely be a few strange folk at the 'edges' of the internet who will have problems and whine.
What percentage of those strange folk are the strange folk who have problems and whine even when things are operating correctly?
participants (6)
-
Adam Clark
-
Jared Mauch
-
Joe Provo
-
Philip Smith
-
Randy Bush
-
Valdis.Kletnieks@vt.edu