This Ready Solution can co-exist with other real-time video analytics solutions. Multiple solutions can be deployed on a single edge node and take advantage of the compute power that is available in the PowerEdge server. These solutions can be installed either as virtual machines to provide isolation or as containers.

Co-existing with Ready Solution for AI: Retail Loss Prevention 

As an example, this ready solution can co-exist with Ready Solutions for AI: Retail Loss Prevention. A retail store can deploy both these solutions for retail loss prevention and video analytics for vehicle identification and customer demographics. A single PowerEdge server can be used as the edge node to support such a scenario.

We recommend that you dedicate NVIDIA T4 GPUs for each of the use cases to provide predictable performance. For example, the PowerEdge R7525 server can be configured with four GPUs, as shown in the following figure:

Figure 5.              Retail loss prevention and video analytics solution

You can dedicate two GPUs for retail loss prevention supporting eight to nine self-service check-out terminals and two GPUs for vehicle identification and customer demographics supporting approximately 15 camera streams.