Convergence with Cisco OER

[maheick]

We are in the process of designing a fully redundant topology that requires automatic failover. Cisco OER seems to be a good fit however there is one requirement of the design that calls for a fast convergence. It appears as though Cisco OER has a minimum peridic timer of 3 minutes.

However, one of the main purposes of this design is to provide connectivity between our two data centers and more specifically it requires that failover be detected within 2-3 minutes. If detection takes closer to 3 minutes, we run the risk of having to go through a very manaully intensive process to regaining sync.

Internally we run EIGRP as our routing protocol. Does OER work with EIGRP to determine link availbility or are the two processes independent? What other strategies could be incorporated into the design to speed up convergence?

Thanks in advance.

 

[brianinms]

OER and EIGRP are to separate processes. Eigrp should be enough to balance between two connections, however OER uses IP SLA to determine the better connection.

What are your 2 connections going to be? If they are point to point T1s you could always use CEF.

 

[maheick]

Both links are PTP with two different providers. I am more worried about the brown conditions more than anything. Something that EIGRP or any other routing protocol for that matter could be prone to.

However I think I am too overly concerned. We have a decent QoS policy to manage congestion and don't foresee congestion as even an issue.

I have decided against using OER on the Enterpise because of the potential convergence.

Thanks again for your input.

 

[brianinms]

If they are true Point-to-point T1's that you can use cef and load balance between them. In addition you can create a Multilink group and bond the 2 T1s together. Even if you lost 1 T1 you would still have 1.5 mb.