Cloud Spanner

  • Cloud Spanner is a highly scalable cloud based relational database
  • Scales to petabytes with strong consistency and high availability
  • Globally distributed with automated sharding and transaction processing
  • Used for financial and inventory applications
  • Benefits over relational databases
    • High Availability
    • Horizonal Scalability
    • Automatic Replication
  • Consider Cloud Spanner if
    • Reached scaling limits of an existing SQL database system
    • Require transactional consistency
    • Need to a globally distributed database with strong consistency
  • A Cloud Spanner instance is
    • Replicated over a number of GCP zones e.g. us-central1, europe-west1
    • Zones can be within or across regions
    • Each zone in the instance contains a full replica of the database
Cloud Spanner Instance
  • node-count N determines the number of nodes allocated to an instance
    • Number of service tasks = N
    • Each node has 2TB of storage