Open
Description
Please add your learnings and results here about your flatbuffer investigations, so that we can have a detailed discussion in the performance and packaging WG.
Pros:
- faster than protobuffers -> especially on sensor model side (de-serialization)
Cons:
- more complex code especially for serialization -> Are the improvement gains worth the refactoring effort for existing code?
General remarks:
- Are there other possibilties to acchieve better performance?
- Flexibility vs performance: Do we want to keep OSI as flexible as it is right now and life with performance limitations or not?