总结

在世界各地拥有多台服务器是最佳选择

• 提高availability

  ：确保如果一台服务器发生故障，其他服务器仍可以处理请求，从而降低停机风险。

• 增强reliability

  ：在多台服务器上分配负载，降低单点故障的可能性。

• 提高fault tolerance

  ：即使某些组件出现故障，系统仍能继续运行。

• 减少latency

  ：将服务器放置在更靠近用户的位置，减少数据传输所需的时间。

• 提高throughput

  ：通过分配负载，使系统能够同时处理更多请求和数据。

Intro

• Key Focus: Thought process and trade-offs, not memorizing facts.
• Core Operations:
  • Move data
  • Store data
  • Transform data

Data Movement

• Within a single machine: From RAM to CPU and disk.
• Across machines: Within the same data center or globally.

Data Storage - several ways, can choose what you want

• Data storage varies: RAM (volatile) vs. Disk (persistent).
• Different data structures (arrays, binary search trees…) have unique properties and trade-offs.
• In larger systems, data can be stored in databases, blob stores, file systems, and distributed systems, each having its own trade-offs.

Data Transformation

• Aggregating application logs to determine success rates.
• Encoding videos, aggregating views.