Skip to content
Phuc Nguyen edited this page May 26, 2023 · 2 revisions
  • Performance: how system performs in a fixed environment, e.g. performance for 100 user and 2 servers

  • Scalability: how system performs in changing environments, e.g performance when users increasing from 1 to 10 to 100

  • Latency: how long it takes to see the output

  • Throughput: how many inputs can be processed per unit of time

  • CAP: have to support partition tolerance (system works during network failures), make trade-off between consistency (can work with latest data) and availability (can work with some data)

  • Object store: data + metadata + unique identifier, e.g. AWS S3. Good for storing large volume of data and retrieving part of it. Not good for operational data.

Clone this wiki locally