EXPEDITION 02 · 2026
Backend Systems
Cacheline
A cache strategy and load-testing toolkit for a read-heavy product API.
240msp95 latency
68%cache hit rate
2026year explored
The question
Could latency be reduced without hiding stale data or making invalidation impossible to reason about?
The route
I segmented endpoints by consistency needs, introduced bounded TTLs, and made cache behavior visible through structured metrics.
The outcome
The mock system sustained the target load with a smaller database footprint and predictable stale-data windows.
Field note
A cache policy is a product consistency decision disguised as infrastructure.
Instruments
- Java
- Spring Boot
- Redis
- k6