What if a smart GPU-powered camera could share (or even sell) its processing capabilities?

Let's imagine an IoT ecosystem where connected devices can share resources.

For instance, a GPU-powered camera is active 50% of the time. That means it is idle for 50% of the time == 50% of resources are wasted.

What if some other device (i.e. a smart meter) could use the processing capabilities of that camera?

This statement opens a few interesting questions:

* How smart meter could know that the GPU camera provides compute resources?

* Should that GPU camera allow any other device to access its resources, or only authorized devices should be allowed to do so?

* Is there a way that the GPU camera could benefit from sharing its compute resources?

In the simplest design, all devices would be connected to the same backend.

That backed would manage the device registry (including capabilities and real-time status of all devices).

In this solution, a smart meter queries the backend for available compute resources. As a result, it receives a list of available devices with respective temporary credentials.

The smart meter can decide which resource (device) to consume. It directly connects to the GPU camera to run some analytics on gathered data and receives the output.

Used temporary credentials are tied to the identifiers of the consumer (smart meter) and the provider (GPU camera), so it is possible to track the resource usage and reward the camera's owner.

I think that this approach enables interesting business cases and optimizes resources (less idle resources would be wasted).

Smart cities are natural ecosystems where sharing/trading free capacity would be very beneficial.

I am aware that this idea is extremely simplified and there would be many security/privacy/interoperability issues, yet it looks achievable to me.

What do you think?