Google Extends Their Services with Cloud SQL
Google is making MySQL available in the cloud as a fully managed service, including a JSON API for programmatic management.
Google is expanding their cloud services with yet another addition, Cloud SQL. This “fully managed” web service offers a a scalable MySQL database running on Google’s infrastructure. Customers are given several ways to connect to a MySQL instance in the cloud: command prompt, API console, admin and reporting tools, external applications, GAE Java & Python applications, and GAE scripts. Instances can be imported/exported with mysqldump.
Google is also making available an experimental JSON API, a RESTful interface providing programmatic administrative capabilities for MySQL: creating, deleting, restarting, restoring from backup, importing/exporting and getting various information about instances.
Although Google provides both command line and API administration capabilities, MySQL instances are automatically managed, being synchronously or asynchronously replicated across different geographic locations, scaled up/down as needed, backup and patched as required.
There are some limitations: user-defined functions and native MySQL replication are not supported. Also a number of MySQL statements are not supported (PDF): INSTALL/UNINSTALL PLUGIN, CREATE FUNCTION, and a few others.
On the heels of Google’s announcement, Amazon reminded customers that AWS RDS has provided MySQL in the cloud since 2009. They have also offered Oracle Database and Microsoft SQL Server since 2012. Amazon has finally decided to take AWS RDS out of beta, announcing general availability and a SLA with “99.95% availability for Multi-AZ database instances on a monthly basis.”
A direct comparison of pricing shows that AWS RDS is cheaper than corresponding Google Cloud SQL pay-per-hour options, but one needs to consider other costs such as data storage and transfer fees, etc. We can say though that Google Cloud SQL supports instances up to 100GB in size while Amazon RDS goes up to 3TB. Also, Google does not offer a SLA.
None of the two companies provides Postgres in the cloud. This can be obtain through third parties, such as Heroku which offers Postgres as a service on AWS.