Chapter Assessment: Cost, Latency, and Scale Patterns

This review covers Cost, Latency, and Scale Patterns, from the token arithmetic of one loop to the reduced bill in Desk's case study. The bank holds 24 questions, and each attempt draws ten, so a second attempt tests a different mix. Most items ask you to price a design or choose a change, and many require arithmetic rather than recall. Have three things ready before you start. Step n sends 6,000 + 1,800(n - 1) tokens, so the total sent grows with the square of the run length. A cache read costs 0.1 units and a five-minute write costs 1.25 units. Each model keeps its own cache