Choosing Between Push and Pull Models for High-Scale Social Feed Generation
Analyze the trade-offs between Push and Pull models for social feed generation, focusing on write amplification and the 'celebrity problem' in distributed systems.
ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Analyze the trade-offs between Push and Pull models for social feed generation, focusing on write amplification and the 'celebrity problem' in distributed systems.
LeetCode's evaluation pipeline uses a container-per-submission architecture with cgroup-enforced resource limits, output comparison, and sandbox isolation. Understanding the trust boundaries and failure modes explains common WA/TLE/RE verdicts that pass locally.
Stop guessing who owns which task in your process maps. Learn how UML 2.5 Activity Partitions (swimlanes) clarify system responsibilities and handle concurrency.
An architecture note on Kdeneon's manifest-based plugin system: trust boundaries at manifest validation, signature verification checks, rendering fallback behavior, and the failure modes that shape upgrades.
Learn how to use McRouter to scale distributed Memcached clusters, prevent cache stampedes using leases, and implement consistent hashing for high-read workloads.
Learn how to implement the Fan-out pattern to solve timeline scalability issues, moving from expensive database joins to pre-computed cached feeds.
Learn how to solve the 'Celebrity Problem' in real-time feeds by implementing a hybrid Fan-out-on-Write and Fan-out-on-Load architecture to balance latency and write amplification.
Learn how the fan-out pattern solves the read-latency problem in social feeds and why a hybrid push/pull model is necessary to handle celebrity accounts and write amplification.
Learn how to build a scalable reputation system by decoupling high-frequency voting actions from user profile updates using asynchronous events and atomic increments.