The Path

16 mental models · 4 scenarios

NOVICEfoundations

STUDIEDscaling

III

EXPERTadvanced

PHASE 1

PHASE 2

PHASE 3

01currently studyingData Modeling & Schema DesignEvery system is just entities and relationships. Before you think about scale, think about what you're storing and why.Read the fundamentals →

Design a Yogurt Ordering System

Design a User Profile and Authentication Schema

02upcomingAPI DesignAn API is a contract between you and your consumers. Design for the consumer, not the implementation.Read the fundamentals →

03upcomingRelational DatabasesTables are truth. Indexes are shortcuts. Joins are the price you pay for normalization.Read the fundamentals →

04upcomingCaching FundamentalsA cache is a bet — you're trading consistency for speed. Know what you're willing to lose.Read the fundamentals →

05upcomingScalability FundamentalsStateless services scale horizontally. State is the hard part.Read the fundamentals →

06upcomingDatabase ScalingYou scale reads first (replicas), then writes (sharding). Never shard before you have to.Read the fundamentals →

07upcomingConsistent HashingWhen your hash space is fixed, adding or removing nodes reshuffles every key. Consistent hashing limits that chaos to neighbors only.Read the fundamentals →

08upcomingMessage Queues & Async ProcessingA queue decouples the producer from the consumer. You send a message and forget — the queue handles delivery.Read the fundamentals →

09upcomingCAP Theorem & Consistency ModelsWhen the network partitions, you must choose: serve stale data or refuse to answer. There is no third option.Read the fundamentals →

10upcomingRate LimitingRate limiting protects resources from being overwhelmed. You're not rejecting users — you're enforcing fairness.Read the fundamentals →

11upcomingStorage & CDNCDNs move data closer to users. Storage tiers match access frequency. Hot data costs more to store fast.Read the fundamentals →

12upcomingDistributed TransactionsACID is easy on one machine. On multiple machines, you orchestrate it yourself.Read the fundamentals →

13upcomingReal-time SystemsReal-time is a spectrum. Polling is lazy pull. WebSockets is a persistent connection. SSE is a one-way stream.Read the fundamentals →

14upcomingSearch SystemsFull-text search is an inverted index — you look up words to find documents, not documents to find words.Read the fundamentals →

15upcomingObservability & MonitoringYou can't fix what you can't see. Logs are events, metrics are trends, traces are journeys.Read the fundamentals →

16upcomingMicroservices vs MonolithMicroservices trade operational simplicity for deployment flexibility. Only split when the monolith's coupling is hurting you.Read the fundamentals →