Specifications

Best Practices for Virtualizing and Managing Exchange 2013
34
34
VMQ, the interrupts were spread across more processors. However, network load could vary over time. A
fixed number of processors may not have be suitable in all traffic regimes.
Windows Server 2012, on the other hand, dynamically distributes the processing of incoming network
traffic to host processors, based on processor use and network load. In times of heavy network load,
Dynamic VMQ (D-VMQ) automatically uses more processors. In times of light network load, D-VMQ
relinquishes those same processors. This helps to ensure that overall performance is optimized.
As shown in Figure 21, without the VMQ technology and RSS, the majority of network processing burdens
CPU 0 and ultimately limits the scale of the solution. With D-VMQ, processor cores are dynamically
assigned to distribute the workload.
Figure 21: Dynamically distributed workload with D-VMQ for Hyper-V
Best Practices and Recommendations
Some Intel multicore processors may use Intel Hyper-Threading Technology. When Hyper-
Threading Technology is enabled, the actual number of cores that are used by D-VMQ should be
half the total number of logical processors that are available in the system. This is because D-VMQ
spreads the processing across individual physical cores only, and it does not use hyper-threaded
sibling cores.
As an example, if the machine has an Intel processor with four physical cores and Hyper-Threading
Technology is enabled, it will show a total of eight logical processors. However, only four logical
processors are available to VMQ. (VMQ will use cores 0, 2, 4, and 6.)
VMQ provides improved networking performance for the management operating system as a
whole rather than for a specific virtual machine. For best results, treat queues as a scarce,
carefully managed resource. Because queues are allocated to virtual machines on a first-come,
first-served basis, making all virtual machines eligible for a queue may result in some queues being
given to virtual machines with light traffic instead of those with heavier traffic. Enable VMQ only
for those virtual machines with the heaviest inbound traffic. Because VMQ primarily improves
receive-side performance, providing queues for virtual machines that receive the most packets
offers the most benefit for overall performance of the management operating system.