@hackproduct9: CAP theorem, explained in one triangle. 🔺 Most people overcomplicate this. Here's the version that actually helps you in a system design interview. You get three properties — but you can only pick 2: 🔴 Consistency — every read sees the most recent write 🟢 Availability — every request gets an answer 🟠 Partition tolerance — the system keeps working when the network splits The catch: in any real distributed system, network failures WILL happen. So partition tolerance isn't optional. That collapses the whole thing into one real decision: when a partition hits, do you choose consistency or availability? Where the databases land: • C + A → Postgres, MySQL (single node — fine until you scale out) • C + P → MongoDB, HBase (stays correct, may reject requests during a partition) • A + P → Cassandra, DynamoDB (always answers, but reads can be briefly stale) The one question to ask yourself: "Does every read need the most recent write?" Yes → consistency. No → availability. Bring this up early in your next interview — it shapes every decision after it. Which side does your system land on — CP or AP? #HackProduct #systemdesign #databases #coding #interviewprep

HackProduct

Open In TikTok:

Region: US

Tuesday 23 June 2026 00:14:02 GMT