To use Cloud SQL effectively, you should understand some of the concepts onwhich it is built. This page provides an overview of key terms and concepts thatapply to Cloud SQL.

For an introduction to using Cloud SQL, seeCloud SQL overview.

Cloud SQL instance

A Cloud SQL instance corresponds to one virtual machine (VM).The VM includes the database instance and accompanying software containersto keep the database instance up and running.

Database instance

A database instance is the set of software and files that operate the databases:MySQL, PostgreSQL or SQL Server.

Public and private IP

In Cloud SQL, public IP means that the instance is accessible throughthe public internet. In contrast, instances using only private IP are not accessible throughthe public internet, but are accessible through aVirtual Private Cloud (VPC).Cloud SQL instances can have both a public and a privateIP address.

High availability

Cloud SQL instances using high availability (HA) provide greaterreliability than non-HA instances.

HA in Cloud SQL works by having two synchronized instances: a primaryinstance and a standby instance. Each instance has exactly one VM. Each instanceis in a different zone in the same region.

Failover

A failover is when Cloud SQL switches serving from the original primaryinstance to the standby instance.

Autofailover is a mechanism that automatically triggers failover when aCloud SQL instance didn't issue a heartbeat in the previous interval.

Standby instances

The standby instance is used in high availability to replace theprimary instance when failover occurs. The standby instance doesn't appear in theGoogle Cloud console. When failover occurs, connections to the primary instance areautomatically transferred to the standby instance.

Clone

When you clone a Cloud SQL instance, you create a new instance that is acopy of the source instance, but is completely independent. After cloning iscomplete, changes to the source instance are not reflected in the clone, andchanges in the clone are not reflected in the source instance.

Replication

Replication is the ability to create copies of a Cloud SQL instance or anon-premises database, and offload work to the copies. The main reason for usingreplication is to scale the use of data in a database without degradingperformance on the primary instance.

Read replica

The read replica is an exact copy of the primary instance. Data and otherchanges on the primary instance are updated in almost real time on the readreplica. Send your write transactions to the primary instance, and your readrequests to the read replica. The read replica processes queries, read requests,and analytics traffic, thus reducing the load on the primary instance.

Source server

Replication copies transactions from a primary instance to one or more readreplicas. The primary instance is also called the source server. The sourceserver can be a Cloud SQL primary instance, or a server outside ofGoogle Cloud, such as an on-premises server or a server running in adifferent cloud. If the source server is outside of Google Cloud, wecall itReplication from an external server.

Cloud SQL Auth Proxy client

The Cloud SQL Auth Proxy client is open source software maintained by Cloud SQL. Itconnects to a companion process, the Cloud SQL Auth Proxy server, running on yourCloud SQL instance. You run the Cloud SQL Auth Proxy client on your own servers. TheCloud SQL Auth Proxy client can be used to establish a secure SSL/TLS connection to thedatabase instance, and/or to avoid having to open the firewall. Authenticationis done through Identity and Access Management (IAM).