Solve / Performance engineering

2024

Adaptive Range Cache

Ranged-key storage and adaptive cache replacement for large data access patterns.

C++ARCLRULFUPerformance
KEY SIGNAL99% hit rate at 30% cache ratio
Technical visualization for Adaptive Range Cache
01

THE CHALLENGE

What needed to change.

Storing every continuous key independently created avoidable overhead, while fixed eviction policies performed unevenly across workloads.

02

THE APPROACH

How I structured the solution.

I represented continuous key regions as ranges and implemented Adaptive Replacement Cache alongside LRU and LFU.

03

THE OUTCOME

What the work demonstrated.

The design reduced per-range overhead from O(K) to O(1), exceeded 90% hit rate at a 10% cache ratio and reached 99% at 30%.