Architecture Simplified: Google Cloud Bigtable

Join Telegram GCP Study Group

Google Cloud Bigtable is a fully-managed NoSQL database service. It is key-value store type of NoSQL database.
Cloud Bigtable is used for large analytical and operational workloads
Cloud Bigtable is ideal for storing very large amount of data in a key-value store.
Cloud Bigtable provides high read and write throughput at low latency
Cloud Bigtable uses horizontal autoscaling
You can add or reduce more nodes to scale throughput
Cloud Bigtable can store up to Petabytes of data
Cloud Bigtable is suitable data source for MapReduce operations.
Cloud Bigtable is the most suitable for high throughput and scalable applications

When you create Cloud Bigtable instance, you can have up to 4 clusters. Each cluster contains nodes which are computational units.
When you create a table in Cloud Bigtable, it belongs to the Cloud Bigtable instance, not any cluster.
Each cluster is located in a single zone. If you want to create additional clusters, you can do it by creating in other zones in the same region.
When you create a Cloud Bigtable instance, you need to choose if you want to store your data on SSD or HDD. Whatever option you choose, all the clusters and nodes will use the same storage type.
How to choose between SSD and HDD -

Choose HDD for very large dataset, for example >10TB, where latency is not the major concern. You can choose it when data is not frequently accessed. For example - batch workloads or data archival operations.
For latency-sensitive requirement, which requires high performance - you can choose SSD.

If your Cloud Bigtable instance has only one cluster then data is not replicated.
If your application requires high availability and you need to provide multi-cluster data replication, then the number of nodes in each cluster should be the same.

Learn more on Google Cloud Services