4. System Architecture

System Architecture ๋ž€?

System Architecture ๊ฐ€ ์ค‘์š”ํ•œ ์ด์œ 

  • Storage System ์—์„œ ์ž‘๋™ํ•ด์•ผํ•˜๋Š” ์•„ํ‚คํ…์ฒ˜๋Š” ์„ค์ •์ด ๋ณต์žกํ•˜๋‹ค.

  • 'Persistent Component' ๋ฅผ ์œ„ํ•ด์„œ, ์‹œ์Šคํ…œ์€ ์•„๋ž˜ ์š”์†Œ๋“ค์„ ํ•„์š”๋กœํ•ฉ๋‹ˆ๋‹ค.

    • Scalable

    • Load Balancing

    • membership

    • failure detection

    • failure recovery

    • replica synchronization

    • overload handling

    • state transfer

    • concurrency

    • job scheduling

    • request marshalling

    • request routing

    • system monitoring

    • alarming

    • configuration management

์ด Paper ์—์„œ๋Š” Dynamo์—์„œ ์‚ฌ์šฉํ•˜๋Š” ๋ถ„์‚ฐ ์‹œ์Šคํ…œ ๊ธฐ์ˆ (core distributed systems technique) ์— ์ง‘์ค‘ํ•˜์—ฌ ๊ธฐ์ˆ ํ•ฉ๋‹ˆ๋‹ค.

๋ถ„์‚ฐ ์‹œ์Šคํ…œ ๊ธฐ์ˆ ์˜ ์š”์†Œ๋Š” ์•„๋ž˜์™€ ๊ฐ™์Šต๋‹ˆ๋‹ค.

  • Partitioning

  • Replication

  • Versioning

  • Membership

  • Failure handling

  • Scaling

Summary of techniques used in Dynamo

Last updated