Backend Systems: Architectural Patterns
The best way to build intuition for backend architecture is to study how real companies solved real scaling problems. This is a reading list of twelve engineering write-ups we recommend to candidates preparing for backend and system design interviews, with a note on what each one teaches.
Big-Tech Backend Systems
1. How Uber Drives 40M Reads/Sec with Integrated Cache
Uber's integrated caching architecture handles 40 million database reads per second globally. The write-up details the components, data flows, and invalidation strategy of a massive-scale, low-latency distributed cache.
2. Why Netflix Integrated a Service Mesh in Their Backend
Why Netflix adopted a service mesh to manage communication across its microservices: a dedicated layer providing load balancing, service discovery, encryption, and observability, plus the evaluation process and operational benefits.
3. How Uber Computes ETA at Scale
Uber's system for computing accurate arrival times from real-time data across millions of drivers: ingestion pipelines, traffic modeling, map-matching, and the ML models behind the estimate.
4. How Zoom Supports 300 Million Video Calls/Day
The architecture behind Zoom's peak load: media routing, WebRTC, load balancing across global data centers, NAT/firewall traversal, and bandwidth-efficiency optimizations.
The architectural decisions and stack behind Threads: real-time messaging infrastructure, feed ranking, CDNs, and media pipelines, built on lessons from Instagram and WhatsApp.
6. How Pinterest Scaled to 11 Million Users with Only 6 Engineers
Pinterest's early scaling story: AWS services, caching layers, database sharding, and async task queues, and the discipline it takes to grow fast with a six-person team.
7. How Quora Scaled MySQL to 100k+ Queries per Second
Quora's database architecture: sharding strategies, read replicas, query routing, and caching layers tuned for high-throughput, low-latency reads and writes.
8. How Canva Supports 135 Million Monthly Users
How Canva's communication layer, built on RSocket, supports 135 million monthly active users, and the trade-offs versus plain HTTP-based approaches.
9. How LinkedIn Scales to 5 Million Profile Reads per Second
LinkedIn's architecture for read-heavy workloads: profile data storage, caching strategies, and denormalization at massive scale.
10. How Uber Finds Nearby Drivers at 1M Requests per Second
Uber's real-time matching system: spatial indexing, load balancing, and the query optimizations needed to dispatch drivers at a million requests per second.
11. How Instagram Scaled to 14M Users with Only 3 Engineers
Instagram's famously lean early stack, and the coding practices and caching strategies that let three engineers support fourteen million users.
12. Prime Video: Amazon's Secret to Streaming Video at Scale
The architecture behind Prime Video's global streaming: CDNs, encoding pipelines, DRM, personalization, and fault tolerance.
How Coaches Can Help
Reading case studies builds recognition; interviews demand production. A backend engineering coach helps you close that gap: practicing how to apply these patterns to an unseen problem, pressure-testing your trade-off reasoning, and spotting the gaps in your fundamentals before an interviewer does.
Conclusion
Each of these write-ups is a case study in matching architecture to constraints rather than reaching for the fanciest pattern. Read a couple per week, and for each one ask: what constraint forced this design, and what simpler version did they outgrow? That habit of thinking is what system design interviews actually measure.