Re: 128/9 cite
On 6/8/23 04:16, Randy Bush wrote: thanks aftab i remember a bit more. the hidden command was there to help debug CEF, which was new at the time. the CEFlapods wanted a large blob of prefixes to push the FIB. it kinda pushed the operational FIBs a bit too far :) Was this in lieu of flow cache and process switching, at the time :-)? Mark.
Re: 128/9 cite
On 6/7/23 10:13 PM, Aftab Siddiqui wrote: I definitely read a detailed research paper about that incident long ago but can't find any link with any search keywords. But here is the NANOG archive. https://archive.nanog.org/mailinglist/mailarchives/old_archive/1997-10/msg00095.html https://archive.nanog.org/mailinglist/mailarchives/old_archive/1997-10/msg00110.html https://archive.nanog.org/mailinglist/mailarchives/old_archive/1997-10/msg00108.html Regards, That's the "old" archive. That was merged in the main archive (same one this list uses) a while back. If you sort by date you'll see a number of threads talking about the 70k BGP table size (lol). It's a fun read to see where we've come from in 26 years. https://mailman.nanog.org/pipermail/nanog/1997-October/date.html#123950 -- Bryan Fields 727-409-1194 - Voice http://bryanfields.net
Re: 128/9 cite
thanks aftab i remember a bit more. the hidden command was there to help debug CEF, which was new at the time. the CEFlapods wanted a large blob of prefixes to push the FIB. it kinda pushed the operational FIBs a bit too far :) randy
Re: 128/9 cite
On Thu, 8 Jun 2023 at 02:15, Randy Bush wrote: > doug madory is asking me for a cite for the exciting 1997/8 128/9 bgp > event. my memory as reported to doug is > > soon after the 7007 incident, an engineer in a UUNET lab, not > realizing they were connected to the real internet, used the hidden > bgp test command to generate 128/9 chopped into /24s. took uunet > down, but not before it propagated. > > does anyone have a useful cite? > I definitely read a detailed research paper about that incident long ago but can't find any link with any search keywords. But here is the NANOG archive. https://archive.nanog.org/mailinglist/mailarchives/old_archive/1997-10/msg00095.html https://archive.nanog.org/mailinglist/mailarchives/old_archive/1997-10/msg00110.html https://archive.nanog.org/mailinglist/mailarchives/old_archive/1997-10/msg00108.html Regards, Aftab A. Siddiqui