Presentation: How Netflix Shapes our Fleet for Efficiency and Reliability (opens in new tab)

The speakers explain the inherent tension between service efficiency and reliability at Netflix's global scale. They share a mental model for "risk-adjusted net value," moving beyond simple CPU utilization to focus on capacity buffers. They discuss hardware shaping, proactive traffic steering, and reactive levers like "hammers" and prioritized load shedding to protect critical playback. By Joseph Lynch, Argha C