Exam Ref 40-710: Nic Teaming doubt

Albertopermer

Hello!

On Exam Ref 40-710 book there are a review question related to Nic-Teaming that i don't undestand.

The question is:



Which of the following NIC teaming modes provides fault tolerance and bandwidth aggregation?
A. Hyper-V live migration
B. Switch Independent Mode
C. Switch Dependent Mode
D. Link Aggregation Control Protocol




I think the solution is option D because I had readed in the book the following text:



In Switch Dependent Mode, you can choose static teaming, a generic mode that balances
the traffc between the adapters in the team, or you can opt to use the Link Aggregation
Control Protocol defined in IEEE 802.3ax, assuming that your equipment supports it.




LACP provides fault tolerance and bandwidth aggregation, but the solution guide of the book say:



A. Incorrect: Hyper-V live migration is not a NIC teaming mode.
B. Correct: In Switch Independent Mode, the NICs in the team are connected to
different switches, providing alternate paths through the network.
C. Incorrect: In Switch Dependent Mode, the NICs in the team are connected to the
same switches, providing link aggregation but no fault tolerance.
D. Incorrect: Link Aggregation Control Protocol is not a NIC teaming mode.




It's this solution a mistake or I am wrong?

Thanks for help.

Cisco Inferno

no its right. thats just a trick question i think to get you. LACP is the protocol and not actually a mode of controlling the team. You have to think of what is running the team? Who's in charge? The switch cpu or windows server's cpu?? Either options has its pros and cons.

Switch Dependent
This mode requires the switch to be aware of and to participate in the NIC Teaming. Since the NIC team is dependent on the switch, you must ensure that all of the members of the NIC team are connected to the same physical switch and not spread across multiple different switches. I dont think two switches can control a single team, at least not here. So with this one there is no failover.

Switch Independent
This configuration does not require the switch to participate in the teaming, therefore the network interface adapters may be connected to different switches. But connecting the NICs to different switches is not a requirement, its merely a possibility in this teaming mode. This teaming configuration will also work with any switches, including non-intelligent/team-aware ones, since all of the intelligence required to support NIC Teaming is handled by Windows Server itself. It also adds an additional layer of failover, if more than one switch is utilized.

The difference is whether or not you need to configure the switch to take advantage of link aggregation. You can have a "Switch independent NIC Team" made up of 2x1Gbps NICs that can only ever transport data point-to-point at 1Gbps as it is not distributed across each NIC.,Or you can have a "Switch Independent NIC Team" that uses your switch's Link Aggregation that can transmit data at 2Gbps thanks to LACP, as well as other things like depending on the source and destination MAC address. This works great for any server that has heavy inbound and outbound traffic.

jaso

Unfortunately the answer to the question "why is it like that" is "because the exam reference says so".

In my opinion you are confusing question that is based on theory (which is described on page 22) with what can really be configured in Windows NIC teaming - there really is the LACP listed under Teaming mode (yet as pointed out by Cisco Inferno, it is a protocol and not than a theoretical mode of operation).

If you read the text carefully they are saying it there - Switch independent - different switches (again according to the text, because switch independent could also mean that the teaming is not dependent on the configuration of the switch) with active/active provides also increased throughput. And that is the answer.

In my opinion - should I get this question in a real MS test, I would add a comment and put some defending arguments for my answer. Also the question should be phrased a lot better than that, because from the context of the answer and the theory in the ref book, they want to have fault tolerance on level of physical switches...).