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
Design Guide—Flexible Computer Vision Solutions with Dell APEX Infrastructure
Executive summary Architecture concepts and requirements Solution architecture Validation
Validation overview
Performance results and findings
Conclusion References

Validation overview

Home > Workload Solutions > Computer Vision > Guides > Design Guide—Flexible Computer Vision Solutions with Dell APEX Infrastructure > Validation overview

Validation overview

  • Our design validation used a 4-node cluster for application hosting, and a Dell PowerScale NAS storage device for accumulation and management of video data streamed to a Milestone Systems VMS. Based on recent sizing requests from customers, a common planning ratio for VMS archived streams recorded for historical analysis to "analytics" cameras requiring real-time alerting is approximately 5:1 or 20% real-time. The reason for this ratio is due to the fact that many cameras are static in nature and may not warrant the extra investment required to enable real-time analytics. We simulated a workload with approximately 30% analytics streams to test the upper bounds of typical processing requirements.

    The validation performed on the Dell APEX solution was a subset of what was tested as outlined in the Design Guide—Virtualized Computer Vision for Smart Transportation with Milestone. The goal was to ensure that the per-VM performance of various components (using NVIDIA A16 GPUs) was in line with the metrics obtained with that study.

    Camera simulators

    Due to the nature of testing in a lab environment, it is not always practical to add hundreds or thousands of real cameras to a system during testing. Video camera simulation software is often used for testing safety and security applications, including VMS and computer vision applications. Camera simulation allows system designers to stress test an entire end-to-end CCTV system while controlling complexity and cost. To enable testing, we leveraged a camera simulator provided by Milestone Systems. The videos used were encoded as H.264 and wrapped with a Raw (not MP4) container. Enterprise-class video camera simulators allow testers to specify bandwidth requirements, quality, quantity, encoding, and source video files.

    The camera simulation system used for testing our design needed to meet the following requirements:

    • Camera simulators run on dedicated virtualized computing nodes
    • The proposed bit rate for all cameras is ~3.2 Mb/s.
    • Specific source video files have been used to test the capability of the CV tool features we are targeting.
    • Simulators connect and stream directly to the Milestone XProtect VMS using a dedicated network segment (VLAN).

    Simulator hardware

    All camera simulation was performed on a single Milestone camera simulator virtual machine. The simulator VM simulates 70 cameras with specific video files for testing computer vision applications.