These walkthroughs use the same sequence: define the requirements, sketch the main modules, choose data flows, and scale the bottleneck that measurements expose.
Choose a problem shape
Store and serve user content
Design a paste-sharing service covers short links, object storage, expiration, analytics, and read-heavy traffic.
Fan out a feed and index search
Design a social feed covers timelines, follower fan-out, search indexing, notifications, and celebrity accounts.
Crawl, deduplicate, and search documents
Design a web crawler covers crawl queues, content signatures, freshness, reverse indexes, and bandwidth limits.
Ingest transactions and compute aggregates
Design a personal finance service covers encrypted external credentials, asynchronous extraction, categorization, and batch spending totals.
Traverse a graph spread across machines
Design a social graph covers breadth-first search, shard lookup, cache placement, and reducing machine-to-machine hops.
Cache a hot read path
Design a query cache covers cache-aside reads, LRU eviction, consistent hashing, and cache update rules. Continue with caching.
Rank records from event logs
Design a sales-rank service covers MapReduce aggregation, distributed sorting, time windows, and read-heavy serving.
Scale infrastructure in measured steps
Scale a service on AWS follows a service from one machine through load balancing, replicas, caching, and autoscaling. See load balancing for the shared concepts.
Model a small domain with classes
Object-oriented design covers class relationships, method responsibilities, and data structures for common domain models.
Related foundations
Use the design method to structure a new problem. Review databases, the API design guide, concurrency, and the system design glossary when a walkthrough introduces those topics.