1# Background

The goal of this example is to demonstrate the tradeoffs found in the mobile + cloud architecture, where one has a choice between doing work locally (on the mobile device) versus remotely (in the cloud) with different performance considerations in each case.

This is a very rough initial attempt and still needs significant work.

• Show how CPU-intensive computation can be off-loaded from a mobile app to the cloud, by comparison to a mobile device, an unlimited resource for computation and storage.
• synchronous local foreground tasks in Android (running directly in an event listener)
• asynchronous local background tasks in Android using java.util.concurrent.ExecutorService + Handler (the recommended replacement for the now-deprecated AsyncTask)
• asynchronous access of remote resources (usually RESTful web services) in Android using OkHttp

See also the corresponding web service.

Check out the project using IntelliJ IDEA. This creates the local.properties file with the required line

sdk.dir=<root folder of IntelliJ's Android SDK installation>

In IntelliJ IDEA: Run > Run app

The underlying links invoke the corresponding web service on the number.

• 1013
• 10007
• 100003
• 1000003
• 10000169
• 100000007
• 1000000007

This project originally used loopj's AsyncHttpClient, and this section used to warn against a different, unrelated AsyncHttpClient (the Sonatype/AsyncHttpClient one, which throws android.os.NetworkOnMainThreadException without explicit configuration and drags in an unnecessary slf4j logging dependency) that people sometimes confused it with.

As of the current version, this project no longer uses either: loopj's library has had no real maintenance in years and wraps a repackaged fork of Apache HttpClient (removed from the Android SDK itself at API 23). The remote-checking code now uses OkHttp, the current standard HTTP client for Android/Java, instead, so this historical confusion no longer applies either way.

• Jason Christensen's OOPSLA 2009 presentation