Design a Custom Page-Replacement Policy for a Tier-1 Cloud Provider Simulator
Overview
What this challenge is about.
Use the provided simulator (Python harness wrapping a C++ page-cache model) and the team's 3 anonymized workload traces (web-cache, key-value store, batch analytics). Implement CLOCK-Pro and an LRU-K hybrid that promotes hot pages to DRAM. Benchmark each against the team's existing LRU baseline on hit ratio, average access latency, and tail latency. Plot all three policies across all three workloads. Deliver simulator code, a Jupyter notebook with plots, and a 5-page recommendation memo naming which to prototype, with honest discussion of where each policy degrades.
The Brief
What you'll do, and what you'll demonstrate.
Implement and benchmark two page-replacement policies against production traces and recommend which to prototype for a tiered-memory architecture.
Earning criteria — what you'll demonstrate
- Understand CLOCK-family vs LRU-family page replacement tradeoffs
- Reason about tiered-memory access cost (DRAM vs CXL latency)
- Run rigorous trace-driven benchmarks with confidence intervals
- Write a technical recommendation memo a kernel team can fund
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Career mappings coming soon.