An administrator is creating a vNetwork Standard Switch with virtual machine and VMkernel networking. The administrator has two uplinks attached to the switch but wants to separate the virtual machine and VMKernel traffic to the two uplinks.
Which load-balancing policy will guarantee this requirement is met, even if additional management or virtual machine traffic is added to the switch?
A.
Route based on IP hash
B.
Router based on source MAC hash
C.
Route based on the originating port ID
D.
Use explicit failover
Explanation:
Ref: Page 44 from vsphere-esxi-vcenter-server-50-networking-guide.pdf
Can somebody please explain..??
Question asks how does the traffic gets segregated on the uplinks..and explicit failover order uses one NIC as active and other as standy kinda…
Did i miss something..? plz explain how that is correct answer.
0
0
the explicit failover, it is a non-load balancing option – This method utilizes the failover order to determine which adapters to use for communication. If an active link is unavailable, a link from the standby list is used.
Take look at A.
it uses for transmissions based on source and destination IP address, VM gp and vmk port group are different ips and different tcp/ip sessions as well.
I vote for A.
0
0
D is correct. Explicit failover is a form of load balancing. It’s just that it’s really only used in the event of a failure. You would create two port groups on the vswitch. One for VM traffic, one for vKernel traffic.
In your vKernel pg you would have vmnic0 as active, vmnic1 as standby.
In your VM pg you would have vmnic1 as active and vmnic0 as standby.
That way you protect yourself from a failure in that if either NIC fails, all traffic gets sent over the remaining NIC. But as long as both NIC’s are working, your traffic is kept separate and balanced.
It’s a similar config for creating multipathing for iscsi but that would use two vKernel port groups.
0
0
thank you for the explanation. more like splitting the two uplinks than load-balance, but I looked the question again, it was not asking load-balance, it is “load-balancing policy”, therefore, I am convinced.
0
0