Master Cloud Bigtable's Hierarchical Structure

Cloud Bigtable: Instances, Clusters, and Nodes

• Cloud Bigtable uses a hierarchical structure of instances, clusters, and nodes for storing and managing data.
• Instances are containers for data and can have one or more clusters located in different zones.
• Each cluster has at least one node that manages data and performs maintenance tasks.
• Tables belong to instances, not to clusters or nodes.
• If an instance has more than one cluster, it uses replication, and tables can't be assigned to individual clusters.
• An instance's storage type (SSD or HDD) and application profiles are important properties that must be chosen carefully.
• SSD is often a more efficient and cost-effective choice, but the choice is permanent, and all clusters in an instance must use the same storage type.
• Application profiles control how applications connect to the instance's clusters and provide separate identifiers for each application or function.
• Clusters represent the Bigtable service in a specific location and belong to a single instance.
• An instance can have clusters in up to 8 regions, and each zone can contain only one cluster.
• The number of clusters that can be created in an instance depends on the number of available zones in the chosen regions.
• Each cluster has one or more nodes responsible for managing specific tablets on disk, handling incoming reads and writes, and performing maintenance tasks.