Capacity Estimation: Feed Reads vs Writes
What you'll learn: How to calculate and use the read-to-write ratio to shape your social media feed architecture and caching strategy.
The Core Asymmetry
In a social media feed system, users consume far more content than they create. Think of it like a concert: thousands watch while only a few perform on stage. This creates a massive imbalance between reads (viewing feeds) and writes (creating posts).
Calculating the Ratio
For a typical social media platform:
Write operations:
- User creates 1 post per day on average
- 100M daily active users
- 100M writes/day ≈ 1,157 writes/second
Read operations:
- Each user opens the app 10 times/day
- Each session loads 50 posts
- 100M users × 10 sessions × 50 posts = 50B reads/day ≈ 578,703 reads/second
Read-to-write ratio: 50,000:1 (or more conservatively, 100:1 to 1000:1)
Why This Matters
This extreme skew fundamentally shapes your architecture:
-
Caching is critical — With 99%+ operations being reads, aggressive caching at multiple levels (CDN, application cache, database query cache) becomes non-negotiable for performance and cost.
-
Write optimization is less urgent — You can afford slightly slower, more complex write paths if it makes reads blazingly fast.
-
Resource allocation — Budget far more servers and memory for read replicas and cache layers than for write handling.
-
Pre-computation strategy — The ratio justifies pre-generating feeds asynchronously rather than computing them on-demand.
Key Takeaway: Social media feeds typically exhibit a 100:1 or higher read-to-write ratio, making aggressive caching and read optimization the cornerstone of a scalable architecture.