Sd Algorithm

SD algorithm

This document describes step-by-step framework on how to tackle a system design question

1. Understand the problem and establish Design Scope#

What you need to collect and define:

• Functional Requirements. They may reside in a form of User Stories or Use Cases.
• Edge Cases and Scenarios that will not be covered
• Constraints
• Architectural Characteristics for the whole system or its parts
• Borders of the system under construction

Flag: an agreement with the interviewer on the design. Output (list of artifacts to be obtained): FR (User Stories, Corner Cases), AR+quantitative characteristics

2. Create a High-Level Design#

Sketch your main components/parts/subsystems and the connections between them. Justify, refactor and repeat. If possible, go through a few concrete use cases.

?? Estimations ?? Define Data model (Logical and Physical ERD etc) .

=> maybe add some APIs for Read/Write scenarios for crucial components/parts then define these API and borders.

Flag: an agreement with the interviewer on the design. Output: blueprint of a system under construction. Only main components.

3. Design Deep Dive#

identify and prioritize components in the architecture dig deeper into 2–3 components and select DB types provide different approaches, their pros and cons, and why would you choose one

Flag: does your design meet all functional and non-functional requirements? Output: ??

4. Wrapping up#

Try to answer these follow-up questions: identify the system bottlenecks and discuss potential improvements it could be useful to give the interviewer a recap of your design. you can talk about error cases (network faults, server failure), operation issues (maybe properties of production-ready systems) how to handle next scale curve