Server network redundancy been a hot topic for a while now. It is an ideal situation when we imagine that a server can be connected to multiple switches, with multiple links to provide higher bandwidth, and fault tolerance.
But doing so, creates multiple challenges from both the network and the servers. Simply to put it, there is no protocol that can run between server and switches so that it can be spanning-tree loop free. If we use LACP, to connect two switches, Server will not have any issue, but the network will break with mac address flaps. If we connect the server with LACP to single switch, the bandwidth will increase , but fault tolerance will be lost.
Cisco has their answer to this solution by providing the VSS with Catalyst 6500 where two switches become one, and the Nexus Switches with the VPC.
Switch Fault Tolerance “SFT”
It uses two adapters to connect to to switches. Only the Primary link will be active, and in case of link, adapter failure, second adapter will take the active role. Uses two adapters connected to two switches to provide a fault tolerant network connection in the event that the first adapter, its cabling or the switch fail. Only two adapters can be assigned to an SFT team.
Figure 1: SFT Network Settings
The image shows the basic configuration. Switches will be configured with portfast and LACP in dynamic mode. The switch uplinks will have STP running. When operational, the Etherchannel will have single link being utilized, while second will be on standby.
Providing the virtual switching solution, such as VSS, VPC, is still the best solution for providing high bandwidth and fault tolerance.