How would you handle database normalization versus denormalization in a high-volume logging system for Akamai’s distributed platform?
