RightSignalLearn · track
Systems design
How big services stay fast and up when millions use them at once: the building blocks behind every systems design interview.
0 of 12 doneStart: Latency and throughput
- Latency and throughputThe two ways to measure fast, and why they trade off · 6 min
- CachingKeeping answers close so you don't compute or fetch them twice · 6 min
- Load balancingSpreading requests across many servers · Coming soon
- Consistent hashingAdding a server without reshuffling every key · Coming soon
- ReplicationKeeping copies of data so one failure loses nothing · Coming soon
- ShardingSplitting data across machines when one is not enough · Coming soon
- CAP theoremWhy a split network forces a choice between consistent and available · Coming soon
- Raft consensusHow servers agree on one log even when some fail · Coming soon
- Message queuesLetting parts of a system work at their own pace · Coming soon
- Rate limitingCapping how often a client can call, with a token bucket · Coming soon
- Exponential backoffRetrying politely so a struggling service can recover · 3 min
- Content delivery networksServing files from a city near the reader · Coming soon