[quagga-dev 12147] Re: Checking in on BGP add_path

Hetu, Seth shetu at akamai.com
Fri Apr 17 15:52:44 BST 2015

Hi Daniel,

Thanks for the link; that's exactly what I was looking for. I'll dig through this and see about adding TX support (I agree that addpath-tx-all-paths is the easiest place to start, and that's probably all I'll need for the moment.)

It'll probably be about a month before I start on coding; I'll definitely ping you if I come up with any questions on add_path. Thanks again!


From: Daniel Walton <dwalton at cumulusnetworks.com<mailto:dwalton at cumulusnetworks.com>>
Date: Friday, April 17, 2015 at 10:43 AM
To: Seth Hetu <shetu at akamai.com<mailto:shetu at akamai.com>>
Cc: "quagga-dev at lists.quagga.net<mailto:quagga-dev at lists.quagga.net>" <quagga-dev at lists.quagga.net<mailto:quagga-dev at lists.quagga.net>>
Subject: Re: [quagga-dev 12145] Checking in on BGP add_path

Hi Seth,
I added support for RX but have not had time to work on TX.  The patch for RX is here:

We just pushed this to our github repo about 5 minutes ago so none of the maintainers have looked at this yet.

If you want to tackle TX support that would be great :)  One thing to think about is that by default when you implement addpath TX, BGP will not send any additional paths.  You need to implement some feature that uses addpath that determines what additional paths to TX.  Some examples:

  *   neighbor x.x.x.x addpath-tx-all-paths - Maybe you have some collector that you want to feed all of the routes to
  *   neighbor x.x.x.x addpath-tx-bestpath-per-nexthop - Handy for RR scenario where you want to let all of the RRCs know one path for each nexthop
  *   neighbor x.x.x.x addpath-tx-bestpath-per-neighbor-as - Can be used to stop MED oscillation

The "all-paths" one is probably the easiest to start with as the latter ones require changes to bestpath to make it calculate the bestpath per nexthop and bestpath per neighbor-as.

I am one of the authors of the add-path draft so feel free to ping me if you have any questions.


On Fri, Apr 17, 2015 at 10:07 AM, Hetu, Seth <shetu at akamai.com<mailto:shetu at akamai.com>> wrote:
Good morning all,

I just wanted to get a discussion started on the BGP add_path extension (https://tools.ietf.org/html/draft-ietf-idr-add-paths-10<https://urldefense.proofpoint.com/v2/url?u=https-3A__tools.ietf.org_html_draft-2Dietf-2Didr-2Dadd-2Dpaths-2D10&d=AwMFaQ&c=96ZbZZcaMF4w0F4jpN6LZg&r=dFtjB4CNUTozJq_F3IVPgQ&m=Iz4wX8h30haLCSJPuQzqQq5fuKmyR2L8XHbL_cRvdOI&s=juhLqS-80_Q1BqT390k4TIN69Q7u7lcTo8X-PNKBxWg&e=>).

In particular, I found a Quagga-dev email from late 2014 that states Quagga does not currently support it, and was wondering if anyone is planning on working on this, or has a test implementation (in a branch somewhere) I can look at.

If not, this is something I would be interested in adding some time over the next few months, as it relates directly to my job. I've got approval from my employer to open source all changes, so I figured I'd contact the Quagga list first to see what the preferred way is to make this change. As I understand it, the change is simple but quite pervasive, as it affects the storage format of many message types.

If this is not something you guys are interested in, that's fine too. Either way, would like to hear your thoughts.


Quagga-dev mailing list
Quagga-dev at lists.quagga.net<mailto:Quagga-dev at lists.quagga.net>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quagga.net/pipermail/quagga-dev/attachments/20150417/1df19137/attachment-0001.html>

More information about the Quagga-dev mailing list