Scaling Writes

Patterns # Scaling Writes Learn about how to scale writes in your system design interview.Scaling WritesPublished Jul 10, 2025 Premium users can view this video once signed in **📈 Scaling Writes** addresses the challenge of handling high-volume write operations when a single database or single server becomes the bottleneck. As your application grows from hundreds to millions of writes per second, individual components hit hard limits on disk I/O, CPU, and network bandwidth and interviewers love to probe these bottlenecks. ## The Challenge Many system design problems start with modest scaling requirements before the interviewer throws down the gauntlet: "how does it scale?" While you might be familiar with tools to handle the read side of the equation (e.g. read replicas, caching, etc.) the write side is often a much bigger challenge. Bursty, high-throughput writes with lots of contention can be a nightmare to build around and there are a bunch of different choices you can make to handle them or make them worse. Interviewers love to probe bottlenecks in your solution to see how you would react to the runaway success of their next product (whether that's realistic or not is a separate discussion!). Write Challenges In this pattern we're going to walk through the various scenarios and challenges you should expect to see in a system design interview, and talk about the strategies you can use to scale your system. Problem Breakdowns with Scaling Writes Pattern YouTube Top K Strava Rate Limiter Ad Click Aggregator FB Post Search Metrics Monitoring ## The Solution Write scaling isn't (only) about throwing more hardware at the problem, there's a bunch of architectural choices we can make which improve the system's ability to scale. A combination of four strategies will allow you to scale writes beyond what a single, unoptimized database or server can handle: * Vertical Scaling and Database Choices * Sharding and Partitioning * Handling Bursts with Queues and Load Shedding * Batching and Hierarchical Aggregation Let's first talk about how we can scale while staying safely in a single-server, single-database architecture before we start to throw more hardware at the problem! ### Vertical Scaling and Write Optimization Vertical ScalingDatabase Choices ### Sharding and Partitioning Horizontal ShardingVertical Partitioning ### Handling Bursts with Queues and Load Shedding Write Queues for Burst HandlingLoad Shedding Strategies ### Batching and Hierarchical Aggregation BatchingHierarchical Aggregation ## When to Use in Interviews ### Common Interview Scenarios ### When NOT to Use in Interviews ## Common Deep Dives ### "How do you handle resharding when you need to add more shards?" ### "What happens when you have a hot key that's too popular for even a single shard?" Split All KeysSplit Hot Keys Dynamically ## Conclusion