You could split your announcement to balance out the incoming traffic.
You could see our announcement at looking glass and query
bgp regexp _4787_
It workout just nicely on the incomming traffic look at:
http://noc.cbn.net.id/bandwidth.html
For outgoing we use policy routing.
On 16 Sep 98, at 20:43, Jon Lewis wrote:
>
> It appears to do what I'd intended. What makes you say that? With a pair
> of T1's to 2 different providers, I'm balancing output traffic by using 2
> route-maps that uses a single as-path access-list to selectively send
> traffic for certain paths through one provider, and send everything else
> to the other. That part's working fine, and the traffic balances quite
> nicely. The real problem (and I don't know how/if I can solve it) is
> input traffic.
>
> With 1 T1 to a tier 1 backbone, and 1 to a regional that connects to
> several tier 1's, most of our input is coming from the regional because
> one of their tier 1 providers connects to much more of the net than our
> tier 1. I do prepending on our announcement to our tier 1 to make the
> as-paths equal in length...but we still aren't getting much input on that
> T1. I'm afraid if I take out the prepend, nearly all our traffic will
> come in that T1...but I may try it and see. For anyone interested, sh ip
> bgp 209.212.128.0 will give you any details I've omitted.
>
Tony S. Hariman
http://www.tsh.or.id
Tel: +62(21)574-2488
tonyha@compuserve.com
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:18 EDT