You describe the what instead of the how. Instead of asking for “10 cores” (which may not deliver the performance you expect), you can ask to try to achieve a latency objective — a goal that’s portable.

Now resource providers can optimize for your desired outcomes rather than rigid configurations.