From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Sun Jul 11 2004 - 01:11:23 GMT-3
At 9:01 PM -0700 7/10/04, Joe Deleonardo wrote:
>Hi Robbie
>
>That was kind of my original question... by the way thanks Howard.
>
>If you configure the prefix list on one end it automagically by way of the
>ORF config, tells the sending router what should be sent.
Right. Incidentally, the IETF is working on specifications to do it
by prefix list, AS path, and community. Expect all of these soon at
your local router!
Sue Hares and I have had a running argument for years that, as one of
the ORF inventors, she finally won. I've claimed BGP doesn't actually
send policy -- it sends the information for policy rules inside the
router to make decisions.
ORF, however, actually sends a policy to a peer.
>
>
>It's a pretty cool feature, saving CPU cycles on both routers, the only
>problem in the real world is that you have to have the eBGP peers'
>cooperation.
It's being adopted widely among ISPs.
>
>Not a problem in the lab though. :)
>
>Have a good one,
>
>Joe
>
>
>-----Original Message-----
>From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
>robbie
>Sent: Saturday, July 10, 2004 4:36 PM
>To: ccielab@groupstudy.com
>Subject: Re: BGP ORF
>
>I'm just now reading about the wonderful world that is ORF - something
>that I'm not at all familiar with. In the article mentioned, most of it
>makes sense, except for how the prefix-list 'FILTER' is applied to the
>ORF peering arrangement - can anyone clarify that for me? It doesn't
>seem too intuitive that one would create a prefix-list that's just
>automagically applied to the ORF instance in the address-family
>configuration without being mentioned.
>
>Thanks in advance,
>Robbie
>
>Howard C. Berkowitz wrote:
>
>> At 2:30 PM -0700 7/10/04, Joe Deleonardo wrote:
>>
>>>
>http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120
>
>>>
>>> limit/120st/120st11/bgpbporf.htm
>>>
>>> Does anyone understand what's really going on here. At first glance it
>>> seems
>>> like a filter would do the same job.
>>>
>>> When I read into it, it almost seems like a peer that wants to accept
>>> limited routes from an eBGP peer sends a message that says "only send me
>>> these prefixes." Otherwise, if that's not the case, I don't see the
>>> benefit
>>> of doing this. You might as well use a regular filter.
>>>
>>> Can anyone confirm my suspicions about what's going on with this feature?
>>
>>
>> Sure. ORF causes your peer router to block routes it would otherwise
>> transmit and you would reject. At the first level, this saves bandwidth
>> that would be otherwise consumed by your inbound filter.
>>
>> When ORF is implemented widely, it lowers the overall filtering load on
> > all routers, since only desired traffic will be received.
This archive was generated by hypermail 2.1.4 : Sun Aug 01 2004 - 10:11:51 GMT-3