How can I integrate the deepstream SDK output for further processing in another machine

I am implementing a model that need to process one more level of processing before pushing the results to end service.
I can think of two approaches

  1. Doing the secondary processing inside the jetson device and giving final output to next stage- Needs deepstream integration with another python program inside the same device.
  2. Performing Secondary Processing in another system-Needs an interface mechanism to connect with another system.

I would like to know, pros and cons of each approach and whis is the better approach for realtime processing.

• Hardware Platform (Jetson)
• DeepStream Version 5.0.1
• JetPack Version (4.4)

Your question is too general to be answered.