I think the later proposal with random values per message are not that easy to implement. A normal distribution is easy. However to have more realistic benchmarks one would model their message distribution. Implementing different random distributions is not that easy.
Implement original proposal of #14.
I think the later proposal with random values per message are not that easy to implement. A normal distribution is easy. However to have more realistic benchmarks one would model their message distribution. Implementing different random distributions is not that easy.