Encryption of onnx files #24
Labels
enhancement
New feature or request
question
Further information is requested
wontfix
This will not be worked on
When creating FMUs using this tool the resource folder in the FMU will contain an onnx file that describes the ML Model contained in the FMU. This means that if someone has access to the FMU file they aslo have access to an full explaination of the ML Model that can easily be parsed and read using a fitting tool.
If we want to be able to share these FMUs without exposing the ML Model inside we need to do something about the onnx file.
One thing that could be done is encrypting the file itself and decrypt it when the FMU is in use. This could work by having the encryption key "baked into" the binary so that it is not easily readable/accessible from the FMU.
This could work by having the encryption key in a .h file that is outside the folder containing the fmu source, but is compiled into the binary by linking it as a library in the cmake file.
It is still unsure if this is something that is worth implementing or if it is even needed by any use case.
The text was updated successfully, but these errors were encountered: