balancing policy for irtual Distributed Switch

[TheVMinator]

I am using a blade chassis where all traffic is being routed through two 40GB uplinks, including NFS, production VM traffic, management traffic.

What load balancing policy should be used on my virtual distributed switch? (route based on originating port ID, source mac hash, IP hash, route based on physical NIC load).  I know route based on physical NIC load is the only one that can load balance NFS traffic.  It requires Enterprise Plus and vDS, which are both in place.  

Should I use load based teaming for NFS port groups and route-based-on-originating port ID for the port groups of other traffic?

[BenLoveday]

Hi there,

I would suggest using Route based on physical nic load for all port groups, particularly in your case with such massive uplink bandwidth :smileysilly:. I suppose this probably won't balance the load evenly across both uplinks but with large bandwidth uplinks this might not be that much of an issue compared with say multiple 1Gb ones.

Hope this helps!

Cheers,

Ben Loveday