Optimizing Traffic with HSRP and OTV

AB.

The scenario described in the question is as follows: two data centers are interconnected using the OTV overlay protocol, and HSRP is configured to provide first-hop redundancy for the servers in each data center. However, traffic between two ESXi virtual hosts on the same site is routed suboptimally through the OTV overlay, which means that it takes a longer path than necessary to reach its destination.

To optimize the traffic, we need to understand why it is being routed suboptimally in the first place. One possible reason is that the HSRP active router for the virtual hosts' subnet is located in the other data center, and traffic is being sent there first before being forwarded to the destination host. This is because HSRP operates at the network layer (Layer 3) and does not take into account the location of the virtual hosts at the data link layer (Layer 2).

To optimize the traffic, we can take two actions:

1. Disable first-hop redundancy: This is not a recommended solution because it eliminates the redundancy provided by HSRP and may lead to a single point of failure.

2. Filter HSRP traffic by using a Layer 3 VACL on the OTV edge devices or a Layer 2 MAC list on the ESXi vSwitch: This solution filters HSRP packets at the appropriate layer, allowing the virtual hosts to communicate directly with each other without going through the HSRP active router. This can be done at either the network layer (Layer 3) or the data link layer (Layer 2) depending on the specific implementation. Both options (B and E) are valid solutions to the problem.

In summary, the best solution to optimize traffic between two ESXi virtual hosts on the same site when HSRP is configured between two data centers interconnected with OTV is to filter HSRP traffic by using a Layer 3 VACL on the OTV edge devices or a Layer 2 MAC list on the ESXi vSwitch.