Hi, I have a question that so far I haven't been able to find a suitable answer for.
This is focused from an ISP perspective.
So suppose I have the following scenario:
I have a BGP transit area. On each edge of the my network I have a eBGP connection to the same client for redundancy. This client has his own ASN, iBGP and prefixes. I'm receiving the same NLRI from the client through both sides. Let's assume he's advertzing the prefix 10.10.0.0/16 through both ends. I'm receiving it with no problems and I'm passing it along to the next providers with whom I also have multiple ingress and egress points.
Something like this:
PROVIDER A PROVIDER A
| |
eBGP eBGP
| |
CLIENT A (ASN65100) --- eBGP --- MYROUTERA(ASN65200) ----- iBGP --- MYROUTERB(ASN65200) --- eBGP --- CLIENT A(ASN65100)
| |
eBGP eBGP
| |
PROVIDER B PROVIDER B
Let's say my client pays for a 10Mb. Both links are configured to 10Mb so that each can handle the load in case the other one fails and both are always active.
So my question is:
How can I shaped or police the client's traffic across multiple points of entry on different routers so that it won't go beyond the 10Mb. The same scenario applies on how can I limit traffic coming from the providers A, and B destined to the client's prefix: 10.10.0.0/16.
I don't mean using MED, local-pref, weight. Sure I can funnel all the traffic through one single point, but consider that I'm also trying to move away from basic routing and more into PfR, which mean that I have more granular control of the flows. Perhaps there is a PfR service-policy or something that can help out.
For this scenario I'm using 7200 as my routers. If there is a solution that assumes any other model don't hesitate to post it.
TLDR; How can I police or shape across multi interfaces on different routers?
Thanks a lot for any pointers or help.
