4 ms·

This is great. For many high velocity low runtime workloads this is going to be a result in a significant cost savings.