3. Spanner introduction
● Global Scale
Horizontally scalable across regions and data centers, from 1 to hundreds or thousands of nodes
● Fully Managed
Database management service means replication and maintenance are automatic
● Relational Semantics
Everything you would expect from a relational database - schemas, ACID transactions, and SQL queries (ANSI
2011)
● Multi Language Support
Client libraries in Java, Go, Python, Node.js and more coming. JDBC driver for connectivity with popular third-party
tools
4. Spanner introduction
● Transactional Consistency
Purpose built for strong, global transactional consistency
● Enterprise Grade Security
Data-layer encryption, IAM integration for access and controls, and audit logging
● Highly Available
Whenever, wherever, your data is highly available
6. Create new Spanner instance
Performance guidance
● Each Spanner node can provide up to 10,000 QPS of reads or 2000 QPS
of writes (writing single rows at 1KB data per row), and 2 TiB storage.
● For optimal performance, we recommend provisioning enough nodes to
keep overall CPU utilization under 75%.
● Minimum of 3 nodes recommended for production environments.
● Note that Cloud Spanner performance is highly dependent on workload,
schema design, and dataset characteristics. The performance numbers
above are estimates, and assume best practices are followed.
16. RESTful result
$ time curl -sS localhost:3000/singer/2 -d firstname=simon -d lastname=su -X POST
$ time time curl -sS localhost:3000/singer/list
See: https://github.com/peihsinsu/try-spaner
17. Summarize
Spanner benefit
● Pay by use with the storage
● Throughput scaleout by extend
instance size
● With transaction and strongly
consistent
● Full managed by Google
~ The most I love ~