Bigtable

• High performance and massively scalable NoSQL database
• What is used for ?
  • High throughput analytic
  • Huge datasets
• Access control
  • Project level or instance level
  • Read/Write/Manage access

Features

• No-ops (Servers)
• An instance: is group of nodes (cluster)
• Auto-scaling
• 02 instances types:
  • Dev: low cost, single node and no replication
  • Prod: replication is available and throughput guarantee

Replicates and changes

• Synchronize date between clusters
• Resizing: add or remove a node without downtime
• Changing disk type (ex: from HDD to SSD) requires new instance

Interacting with Bigtable

• Using cbt tool (recommended)
• With HBase shell

Example

• echo -e "project = [PROJECT_ID]\ninstance = [INSTANCE_ID]" > ~/.cbtrc
• Create table cbt createtable my-table
• Add column family cbt createfamily my-table cf1
• Add to row 1 the value to column c1 cbt set my-table r1 cf1:c1=test-value
• Read table cbt read my-table
• Delete table cbt deletetable my-table

Schema Design ??

Tall and Narrow Tables ??

Key ??

IAM

• Project, instance and table level