Network performance is dependent on application workload and network configuration. Dropped network packets indicate a bottleneck in the network. Slow network performance can be a sign of load-balancing problems.

Problem

Network problems can manifest in many ways:

  • Packets are being dropped.
  • Network latency is high.
  • Data transfer rate is low.

Cause

Network problems can have several causes:

  • Virtual machine network resource shares are too few.
  • Network packet size is too large, which results in high network latency. Use the VMware AppSpeed performance monitoring application or a third-party application to check network latency.
  • Network packet size is too small, which increases the demand for the CPU resources needed for processing each packet. Host CPU, or possibly virtual machine CPU, resources are not enough to handle the load.

Solution

  • Determine whether packets are being dropped by using esxtop or the advanced performance charts to examine the droppedTx and droppedRx network counter values. Verify that VMware Tools is installed on each virtual machine.
  • Check the number of virtual machines assigned to each physical NIC. If necessary, perform load balancing by moving virtual machines to different virtual switches or by adding more NICs to the host. You can also move virtual machines to another host or increase the host CPU or virtual machine CPU.
  • If possible, use vmxnet3 NIC drivers, which are available with VMware Tools. They are optimized for high performance.
  • If virtual machines running on the same host communicate with each other, connect them to the same virtual switch to avoid the cost of transferring packets over the physical network.
  • Assign each physical NIC to a port group and a virtual switch.
  • Use separate physical NICs to handle the different traffic streams, such as network packets generated by virtual machines, iSCSI protocols, vMotion tasks.
  • Ensure that the physical NIC capacity is large enough to handle the network traffic on that virtual switch. If the capacity is not enough, consider using a high-bandwidth physical NIC (10Gbps) or moving some virtual machines to a virtual switch with a lighter load or to a new virtual switch.
  • If packets are being dropped at the virtual switch port, increase the virtual network driver ring buffers where applicable.
  • Verify that the reported speed and duplex settings for the physical NIC match the hardware expectations and that the hardware is configured to run at its maximum capability. For example, verify that 1Gbps NICs are not reset to 100Mbps because they are connected to an older switch.
  • Verify that all NICs are running in full duplex mode. Hardware connectivity issues might result in a NIC resetting itself to a lower speed or half duplex mode.
  • Use vNICs that are TCP Segmentation Offload (TSO)-capable, and verify that TCP Segmentation Offload-Jumbo Frames are enabled where possible.