[c-nsp] Time Warner redundancy with ... RIP / BFD?

[Adam Greene]

Hi,

 

I have a customer who has two upstream providers: us and a T1 provider. The
customer is running eBGP with both providers. They want to drop the T1 and
get Time Warner cable instead as their redundant link (mainly because the
Verizon local loop is so poor). Time Warner doesn't support eBGP on a
standard cable connection, but offered to run RIP with the customer. They
told the customer that they should check with us to see if we support BFD,
to make the redundancy work.

 

My understanding of BFD is that it simply reduces convergence times for
dynamic routing protocols. BFD might be nice in this or any other case, but
I don't see it as essential to what the customer is trying to do.

 

My thought is, much as I don't like the idea, if Time Warner is willing to
run RIP with the customer, the customer could accept a default route from
Time Warner and from us, and can use administrative distance, metrics,
AS-Path prepending, and/or conditional BGP advertisements to achieve the
kind of failover redundancy they are looking for. 

 

Am I missing something? Is BFD actually an integral part of this equation?

 

Thanks,

Adam