Before you configure NIC teaming in Hyper-V, you need to evaluate each option and understand its limitations. Otherwise, you could experience performance issues or even inaccessible workloads. To start, there are basically two ways to configure NIC teaming in conjunction with switches and the greater LAN: switch independent mode and switch dependent mode.
Switch independent mode
The switch independent mode uses multiple switches -- connecting each of the server's teamed network interface card ports to a corresponding port on different switches. For example, if there are two NICs in a team, both NIC ports are connected to a port on two different switches. If there are three NICs in a team, each of the three NIC ports is connected to a port on three different switches and so on. This approach is best for network resilience because a failure in one NIC or switch won't render the corresponding workload unavailable. The switch independent mode can support both active/active and active/standby operation.
The active/active mode allows all of the NIC ports in the team to share the bandwidth. If a failure occurs, the total bandwidth of the team will be reduced by the bandwidth of the failed NIC. For example, if an active/active NIC team uses two 1 GigE NICs, that team can offer up to 2 Gbps to the workload. If one of the NICs fails, the bandwidth would be degraded to 1 Gbps, but the workload would remain available.
The active/standby mode allows all of the traffic to run across a subset of NICs and then fail over to another subset in the team if a failure occurs. For example, if an active/standby NIC team uses two 1 GigE NICs, that team can only offer up to 1 Gbps to the workload. But if one of the NICs fails, the connection would simply fail over to the alternate NIC, and the bandwidth and connectivity would effectively remain unchanged.
Switch dependent mode
The switch dependent mode connects all of the teamed NIC ports to a single network switch. The use of a single switch limits resilience -- if the switch fails, all of the team links are cut off, and the workload becomes inaccessible -- but all of the teamed NICs are aggregated into a single link for best bandwidth. In switch dependent mode, administrators can select static or generic teaming that foregoes the use of a configuration protocol or use the IEEE 802.1ax Link Aggregation Control Protocol to coordinate network traffic between the server and the switch.
When you configure NIC teaming with switch dependent mode, it tries to keep all of the packets of a TCP stream on a single NIC. This helps to keep packets in order and bolster effective performance – out-of-order packets can impair performance -- but this can also limit NIC usage to a single NIC, even if there is more than one NIC in the team. However, multiple workloads can use the same aggregated team.
Navigate the top Hyper-V 2016 features
Focus on intent-driven networking
Dig Deeper on Microsoft Hyper-V management
Related Q&A from Stephen J. Bigelow
Just because software passes functional tests doesn't mean it works. Dig into stress, load, endurance and other performance tests, and their ... Continue Reading
Don't neglect form factor as part of your data center server selection. Instead, figure out what type of environment you need and learn which server ... Continue Reading
Learn how load balancing in the cloud differs from a traditional network traffic distribution, and explore the different services available from AWS,... Continue Reading