Testing and debugging Big Data application is often a tedious process, involving changes of setup, restarts and analysis of log files. In this demo I am gonna present Port, a Live Big Data framework in Pharo. Port can be used to write classical Map/Reduce applications, that will be parallelized by the framework and executed in different running environments (i.e. Pharo images). Port can be deployed simply on your local machine (with an automatised system to spawn different running environments), or on a distributed cluster leveraging on Hadoop YARN.

I will show how developers can write, execute and manage their applications, and the tooling support for testing such applications in different setups. I will also show how the concepts of out-of-place debugging are used to handle failures in the running application.

I am a pre-doctoral researcher at the Vrije Universiteit Brussel (Belgium). After completing my bachelor studies in Italy (Università degli studi di Bologna), I moved to Brussels for a master at the Vrije Universiteit Brussel, in the Software Languages Lab, and decided to start a PhD under the guide of prof. Elisa Gonzalez Boix. I also cooperate with the RMoD research group at INRIA Lille Nord-Europe.

My research focuses on debugging concurrent and parallel applications, particularly Big Data (or data intensive) applications. In this edition of Programming, after having participated in the Student Research Competition last year, I will be presenting the work of my Master Thesis and of my first months as a PhD student. This works focus on out-of-place debugging, a debugging architecture for online live debugging of distributed applications.