Google cloud SQL for production purpose

[Vaishnavi Manjunath]

Hi,

I am developing enterprise level Web application using Google app engine with Java as primary language, also using Google cloud SQL 2nd generation instance as database . My application will have to deal with say half a million users creating lots of bills, invoices , recording transactions on an everyday basis . I did read a lot of posts which said that query caching does not work with cloud SQL and performance is too slow. My application has faster processing needs in terms of CRUD operations and I really need clarity on performance , back up , latency, storage and optimization issues. If at all query caching is not supported with cloud SQL, will enabling query caching and other optimization techniques using JPA providers like eclipselink work ?

Any help asap will be appreciated !

Thanks :)

[George (Cloud Platform Support)]

Hello Vaishnavi,

You are to some extent right, when you express concern about the performance of an app engine application using CloudSQL 2nd Generation: network latency for connections between App Engine standard environment and Second Generation instances is approximately double the latency for connections to First Generation instances. Also, at present, there is no support for query caching.

Besides Cloud SQL, you have the choice of other storage products, that may better suit your purpose.

Regarding scaling, it may be worthwhile bringing to mind that Datastore is designed with scalability in mind, whereas MySQL does not scale well in practice.

This being said, you may design your application from bottom up with scalability in mind, following useful generally applicable good design practices: designing for scale.

You may have a look at other options, such as setting up and managing your own high-performance SQL server instance using the Compute Engine. This way, you can implement query caching as desired at relatively low cost.