freitasm: As for switch involvement, you might have to read this blog post

Zeon:

freitasm: As for switch involvement, you might have to read this blog post

So how would it work over a switch that isn't configured apart from showing the same MAC on both ports?

Just be careful, it might not be all its cracked up to be....

When using Switch Independent mode, there are a few considerations to keep in mind:

• If you are using NIC teaming for Failover, rather than load balancing across multiple active NICs, switch independent mode works great.  Just set your second NIC in the team as the "Standby adapter" and you're all set.  When a failover occurs, the Standby NIC will become active and the server will notify the network switch of this change via standard network protocols.
   
• When using NIC teaming for Load Balancing across multiple active NICs, there's two load balancing modes you can use with "Switch Independent" mode:
   
• Address Hash - Load balances outbound network traffic across all active NICs, but only receivesinbound traffic via one of the NICs in the team.  Since the switch isn't actively involved in the "Switch Independent" teaming mode, it can't load balance inbound traffic across all active NICs.  This mode can work well if your server has little traffic inbound and lots of traffic outbound that you are trying to load balance - Web servers and FTP servers are examples of typical server roles that work well in this scenario.
   
• Hyper-V Port - If your server is a Hyper-V host with multiple running VMs, this load balancing mode is normally preferred in most situations.  When using "Hyper-V Port" load balancing, VM's will be distributed across the network team and each VM's outbound and inbound traffic will be handled by a specific active NIC.  This mode works really well in scenarios where you are consolidating many VM's on a physical Hyper-V host, but where none of the VM's are generating a network load that exceeds the bandwidth of one NIC in the team.  In this use case, NIC teaming provides a very cost-effective way of load balancing the aggregate traffic from all VMs across the active team members, but remember ... each VM is assigned to a specific NIC in the team, so none of the VM's will be able to access more bandwidth than what one NIC provides.