Distributed Database

(Information Science Expert) Lecture on Database

• Distributed Systems

• Utilizing multiple servers

• Consists of a central server and multiple DBMSs

• 2-phase commit

• Cloud databases like BigQuery and DynamoDB also use this

• NoSQL architecture (Do other distributed databases use this too?)

  • Sharding: Dividing data and allocating it to multiple servers
  • Replication: Creating multiple copies of each data and distributing them to multiple servers
    • Replication factor (measure of how much duplication range to have)
    • Trade-off between fault tolerance and storage usage
    • Types: master-slave, peer-to-peer
      • Difference lies in whether there is a “single point of failure”
  • In practice, a combination of the above two is often used.