Cloud Storage with GCP

Cloud Storage

File is stored as Binary large-object (BLOB).

Suitable for backing up and archiving content online.

Files are organized into buckets. A bucket name has to be globally unique. You can choose where to store and serve your data: Multi-region, Dual-Region, Region, and Zones.

These options offer different resiliency and cost.

The other option you can make to your bucket is the storage tier. This option let you reduce storage cost if you don’t need to access data in the bucket frequently.

Permission

You can make uniform rules that apply equally to a group of users. Or you can define fine-grain rules to allow access for specific people.

Data Protection

These are paid addons that increase your data safety.

Soft-delete: Data will be kept for a specified amount of time before being deleted permanently. You can recover data within this period of time.

Versioning: Keeping multiple copies of data.

Transfer Job

You can transfer data from multiple sources into buckets on Google Cloud.

You can either transfer in batch for one-time migration or use Change Data Capture for continuous streaming of new data to your bucket.

Cloud SQL

This service lets you quickly create a SQL database on Google Cloud.

You can have up to:

• 128 CPU cores
• 64TB of storage

in one instance.

Spanner

Spanner is a fully managed relational database. It supports horizontal scaling through Spanner nodes.

This service is helpful when you need global synchronicity/consistency of data.

Firestore

Firestore is a NoSQL cloud database.

Data is stored as documents and therefore does not support SQL. However, Firestore supports transactional behavior for data.

Firestore uses online and offline data synchronization.

Bigtable

Bigtable is a NoSQL that is specialized for massive workloads and high throughput.

This is suitable for analytical applications where you need to read and transform terabytes of data.

Choosing the right storage options

Google’s Guide on Storage Solution