Programming and Deployment of Cloud-based Data Analysis Applications

Cloud Computing constitutes a model capable of enabling the network access in a shared, practical and on demand of different computational resources like networks, memory, application or services. This work has as goal the explanation of the project made within Cloud Computing. After an introduction of the theory that lies behind Cloud computing's technologies, there is the practical part of the the work, starting from a more specific platform, Hadoop, which allows storage and data analysis and then moving to more general purpose platforms, Amazon Web Services and Google App Engine, where different types of services have been tried. The major part of the project is based on Google App Engine, where storage and computational services have been used to run MapReduce jobs. MapReduce is a different programming approach for solving data analysis problems, that is suited for big data. The first jobs are written in python, an imperative programming language. Later on, a functional approach on the same problems has been tried, with the Scala language and Spark platform, to compare the code. As Cloud computing is mainly used to host websites, a simple site was developed as integral part of the work. The development of the site is not explained as it goes beyond this thesis' main focus, only the relevant aspects will be treated.