/24s run amuck... again

Alexander, Nathan M nathan.alexander at qwest.com
Fri Jul 19 18:17:42 UTC 2002


That is of limited usefulness.  1)  They'd all have to have the same next
hop.  2) you would have to hear all three /24's of them from all your peers,
all with the same next hop, metrics, med, etc.  Otherwise the aggregate
route becomes the backup route for other, more specific announcements of the
/24 that you can't aggregate.  It would be nice for the single homed, but
then why are you running bgp?

Nathan Alexander
NMC Shift Manager
703/363-5293(office)
301/442-8695(cell)
877/541-2189(pager)
Nathan.Alexander at qwest.com

- Internal Use Only - Disclose and Distribute only to Qwest Employees and
authorized persons working for Qwest. Disclosure outside of Qwest is
prohibited without authorization.



> -----Original Message-----
> From: Ralph Doncaster [mailto:ralph at istop.com]
> Sent: Friday, July 19, 2002 8:07 AM
> To: Richard A Steenbergen
> Cc: nanog at merit.edu
> Subject: Re: /24s run amuck... again
> 
> 
> 
> On Fri, 19 Jul 2002, Richard A Steenbergen wrote:
> 
> > With all the recent talk about filtering, I figured now was 
> a good time
> > to update my list of evil /24 announcers... There are 
> currently over 63k
> > /24s out of 113k total unfiltered announcements (over 55%).
> 
> Does anyone (Cisco, Juniper, etc.) have an auto-aggregation feature?
> Let's say a router that receives the following prefixes under 
> a single AS
> 206.51.128.0/24
> 206.51.129.0/24
> 206.51.131.0/24
> 
> And the following network is not in the table:
> 206.51.130.0/24
> 
> Then it would be usefull to collapse it into a single 206.52.128/22
> aggregate (especially if the metrics & communities are the same).
> 
> -Ralph
> 
> 



More information about the NANOG mailing list