Performance Engineering: Scenario-Based Questions

Caching improves performance and reduces load by serving data from memory or edge. However, cache invalidation is one of the hardest problems in computer science — getting it wrong can lead to stale data or outages.

🚀 Caching Layers

• Client-side: Browser or mobile app caching (e.g., HTTP cache, localStorage).
• Edge: CDN like Cloudflare, Akamai caches static/dynamic content.
• App/Service Layer: In-memory caches (Redis, Memcached) per service.
• Database Layer: Query result caching, read replicas.

📦 Caching Strategies

• Read-through: App fetches from cache; cache loads if miss.
• Write-through: Updates go to cache and DB simultaneously.
• Write-around: Bypass cache on write, refresh on read miss.
• TTL-based: Time-bound expiry (e.g., 60s freshness window).

🔄 Cache Invalidation Techniques

• Time-based: TTL or sliding expiration.
• Event-based: Invalidate cache on DB update (e.g., pub/sub).
• Manual: Admin flush or cache bust on deploys.
• Versioned keys: Use data versioning in cache key to force refresh.

🧰 Tools

• Redis, Memcached for app-level caching.
• CDNs with edge rules (Cloudflare Workers, Fastly).
• ORM-level cache decorators (e.g., Django cacheops, Spring Cache).

✅ Best Practices

• Use consistent keys and namespacing.
• Monitor hit ratio and eviction patterns.
• Build fallbacks for cache failures or cold starts.
• Document expiration policy per cache.

🚫 Common Pitfalls

• Stale cache served after underlying data changed.
• Hardcoded TTLs or mismatch with data update frequency.
• Cache stampede: simultaneous misses overwhelm DB.

📌 Final Insight

Caching can supercharge performance — but only with strong discipline around consistency and invalidation. Think of cache as a living part of your data pipeline, not a shortcut.