Giuseppe Taibi’s Techlog

Google has launched App Engine, their new scalable computing platform for application developers. Dreams do come true for application developers who now have an incredible array of choices for hosting their apps. With so many choices for hosting web apps on massive clusters of servers, scalability is no longer a problem. The race is still open for ease of deployment, high-availability and manageability. Currently, the runtime environment supports only Python and Django, with more languages to come. The environment features:

• dynamic web serving, with full support for common web technologies
• persistent storage with queries, sorting and transactions
• automatic scaling and load balancing
• APIs for authenticating users and sending email using Google Accounts
• a fully featured local development environment that simulates Google App Engine on your computer

Interestingly, the storage model is not a relational database but rather an object database, as the native ZODB of Zope and Plone. This is the first time that such object database gets exposure to a massive number of developers. It will be interesting to follow the reactions to this choice from the blogosphere.

Google’s App Engine datastore uses a non-relational object model to store information, which allows you to create fast, scalable applications. Because this method of storing data differs from relational databases, such as MySQL, not all functionalities used with relational databases are available with the system. For instance, the Google App Engine datastore does not have the ability to ‘join’ tables that is available in a relational database.

We have full CRUD operation (create, retrieve, update and delete) support, and allow you to query against our datastore using GQL, our query language. You may also use our Query class, which allows you to use filters and ordering when retrieving the data, as well as perform ancestors queries for an entity.

Datastore filters allow you to restrict the data retrieved from the datastore by using common comparison operators: < <= = >= >

The ordering operations allow you to organize how the data is returned from the query. You can view the ordering operations for each type available with the datastore here.

Lastly, when retrieving the data from the datastore, you can specify both a limit and an offset to the results returned. The limit is maximum number of results returned by the query. The fetch offset specifies the number of results to skip when returning the data from the datastore.

See a GQL example at Creating, Getting and Deleting Data.

Get the full scoop at What is Google App Engine.