Demystifying Cloud Spanner multi-region configurations

Опубликовано 16 июня 2021, 16:00

Google Cloud Spanner → goo.gle/3iAn3PJ

Cloud Spanner is a fully managed relational database service that offers unlimited scale, strong consistency and up to 99.999%. How does does the distributed system that is Cloud Spanner work behind the scenes across zone and region barriers to provide all that? Join Neha Deodhar - a Cloud Spanner engineering manager as she explains what happens under the surface.

Chapters:
1:50 - What is Cloud Spanner?
2:48 - Horizontal scalability using splits
3:57 - Replicate data across failure domains
6:28 - Writes in Cloud Spanner
7:17 - Regional instance configuration
9:09 - Multi-region instance configuration
12:08 - Read-write witness replicas
13:47 - Read-only replicas
14:28 - Strong reads
16:11 - Stale reads
17:00 - Default leader region
18:58 - Node vs Replica
20:10 - Leader region failure
22:38 - Non-leader region failure
23:30 - Zone failure
25:04 - Closing remarks

Watch episodes of Spanner: Database Unlimited → goo.gle/SpannerDatabaseUnlimit...
Subscribe to Google Cloud Tech → goo.gle/GoogleCloudTech

#SpannerDatabaseUnlimited

Product: Cloud Spanner; fullname: Neha Deodhar;