Redis Caching Layer: Cutting API Latency at mPokket
Designed and implemented a Redis caching solution that dramatically reduced API response times for a large active user base.
Suraj Nanda
SDE 2 at Acko
From their time as
Software Engineer
mPokket • 2023 - 2025
Overview
At mPokket, frequently accessed data was being fetched from the database on every API call, creating unnecessary latency for a user base of over thirty thousand active users. Suraj identified the opportunity and **initiated a Redis-backed caching solution** to address it.
The Story
At mPokket, frequently accessed data was being fetched from the database on every API call, creating unnecessary latency for a user base of over thirty thousand active users. Suraj identified the opportunity and initiated a Redis-backed caching solution to address it.
He designed the caching layer to intercept reads for high-frequency data, serving responses from Redis rather than hitting the database each time. The implementation also included Redis locks to prevent race conditions on cache writes, ensuring data consistency under concurrent load.
The result was a significant reduction in API latency: from 200ms down to 25ms, an improvement of nearly 90%. The change improved system responsiveness across the platform for the affected user base.