Load Balance and Cache Affinity - Friends, Foes, or Frenemies?

Explore the complex relationship between load balancing and cache affinity in distributed systems through this 83-minute conference talk by Aaron Archer from Google at the Simons Institute. Delve into the fundamental tension between distributing computational load evenly across servers and maintaining data locality for optimal cache performance. Examine how these two critical system design principles can work together, conflict with each other, or create unexpected interactions in large-scale computing environments. Analyze real-world scenarios where load balancing decisions impact cache hit rates and overall system performance. Learn about strategies for managing this trade-off in modern distributed architectures and understand the implications for memory management in high-performance computing systems.