[27341] in North American Network Operators' Group
RE: Cisco - ip verify unicast reverse-path
daemon@ATHENA.MIT.EDU (Paul Ferguson)
Sat Feb 12 21:23:55 2000
Message-Id: <4.2.2.20000212211958.00a5a3d0@lint.cisco.com>
Date: Sat, 12 Feb 2000 21:21:23 -0500
To: jlixfeld@idirect.com
From: Paul Ferguson <ferguson@cisco.com>
Cc: trall@almaden.ibm.com, nanog@merit.edu
In-Reply-To: <594E6BF7D307D311B1F90080C8E25EA357A7B2@MONET>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Errors-To: owner-nanog-outgoing@merit.edu
At 09:18 PM 02/12/2000 -0500, jlixfeld@idirect.com wrote:
>Correct me if I'm wrong, but it has the potential to cause more damage than
>anything else if you network routes asynchronously. I turned it on a while
>ago and noticed my in and outbound traffic dropped by about 60%. I'm sure
>that the 60% drop in traffic had to do with traffic coming into the network
>on one interface and being local pref'd out a different interface.
>
>Because of the asynchronous routing, it's useless in my environment, unless
>there is a knob that doesn't cause it to break under these circumstances.
>
>If not, would this feature not be best suited for ISPs who are not in a BGP
>relationship with their upstreams?
Like everything else, it depends.
There's always RFC2267.
- paul
