Thoughtworks on App Engine For Java

Noch ein Session Video von der Google I/O: ThoughtWorks on App Engine for Java: An Enterprise Cumulonimbus? Präsentiert von Martin Fowler und Rebecca Parsons.

Nicht ganz so interessant wie andere und aus diesem Grund ist eventuell nicht wirklich wert eine ganze Stunde dafür zu opfern. Hier einige Notizen:

Testing

• Local Server clones GAE not yet a 100%. Testing just on GAE is slow. Testing is possible but might be a little arkward until the local server really clones GAE.

Persistence

• GAE uses a Nested Hash Map as a Data Storage. That is perfectly fine but might cause troubles or at least require a new thinking for developers who are used to relational databases.

Concurrency

• Classical Java uses one single memory stack and runs multiple threads in this memory. GAE however only runs one thread per memory. Trying to create a new thread causes an exception. This particularly causes problems for many of the existing Java libraries. Key point: You might not be able to use all Java libraries you’d like to until they’re ported.

How to classify GAE?
This is how they classify GAE:

Quelle: http://code.google.com/intl/de-DE/events/io/sessions/ThoughtWorksAppEngineJava.html

As example for infrastructure clouds they mentioned Amazons EC2. GAE belongs to the Platform category and Salesforce is a common example for Application clouds.

Es gab einige Anmerkungen mehr. Interessierte sollten das Session Video oder die Präsentationsfolien anschauen.