Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
Technical White Paper—Virtual Desktop Infrastructure for the Financial Services Industry
Executive summary VDI design considerations in FSI Solution architecture Standard VDI test results and analysis Conclusion References Appendix A: Definition of performance metrics
Appendix A: Definition of performance metrics

Appendix A: Definition of performance metrics

Home > Workload Solutions > Virtual Desktop Infrastructure > White Papers > Technical White Paper—Virtual Desktop Infrastructure for the Financial Services Industry > Appendix A: Definition of performance metrics

Appendix A: Definition of performance metrics

  • The following table explains the performance metrics used during our testing:

    Table 14. Definition of performance metrics
    Metric Definition
    CPU usage The combined average CPU usage of all compute hosts during the steady-state phase.
    CPU core utilization The CPU utilization percentage of the corresponding core. A core is utilized if one or both of its logical CPUs are used. This figure is averaged across all cores.
    CPU readiness The percentage of time that the VM was ready but could not get a scheduled run on the physical CPU.
    GPU usage The combined average GPU usage of all six installed T4 GPUs during the steady-state phase.
    Consumed memory The average consumed memory across all compute hosts during the steady-state phase.
    Active memory The average amount of guest “physical” memory actively used across all compute hosts during the steady-state phase.
    Network usage Network usage per user is the combined average during the steady state phase divided by the number of users on a host in megabits per second.
    Cluster disk IOPS Read and write Input/Outputs Operations Per Second (IOPS) consumed by all vSAN clients in the cluster, such as virtual machines, stats object, and so on.
    Disk I/O (or cluster) latency Average read and write latency of I/Os generated by all vSAN clients in the cluster, such as virtual machines, stats object, and so on.