Randomized Load Balancer with the Power of Two Choices
Overview
What this challenge is about.
Simulate four placement policies on a 24-hour anonymized connection-establishment trace (around 4.1 billion events): (1) random-1, (2) round-robin, (3) P2C with instantaneous load, (4) P2C with EWMA-smoothed load (alpha = 0.3). Report mean, p99, p99.9, and p99.99 latency for each. Run a theoretical-vs-observed comparison against Mitzenmacher's classical P2C bounds. Deliver a Python simulation harness, a 7-page comparison report with latency CDFs, and a rollout recommendation with a canary plan.
The Brief
What you'll do, and what you'll demonstrate.
Quantify the tail-latency improvement of power-of-two-choices load balancing vs random-1 on a 4.1-billion-event trace and recommend a production rollout policy.
Earning criteria — what you'll demonstrate
- Implement and reason about the Mitzenmacher 'power of two choices' result
- Build a discrete-event simulator that respects real trace inter-arrival times
- Read tail-latency CDFs and connect distribution shape to load-balancing policy
- Translate simulation results into a defensible production rollout plan
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.