Mobile data collection based on Hadoop platform

OData support
Supervisor:
Dr. Ekler Péter
Department of Automation and Applied Informatics

In the recent years mobile devices have gone through such development, that building mobile optimized versions of certain websites has become a very important aspect of web development. However to achieve this goal, developers have to make additional efforts and also need to have the right skills. Because of this, it is a valid expectation from the owners of these sites to be able to measure if they really need these optimizations. Therefore the main purpose of my work is to develop a system which creates a report for site owners about the mobile users of their page.

During the implementation of the system, I was aiming to prepare it to be able to handle a very large set of these sites simultaneously, so I used distributed systems and algorithms. I used the Apache Hadoop software library to accomplish this, since this product has become so popular recently, that it is now one of the most popular BigData solutions. It is important to note that I did not use an actual cluster to test the features of the system, as of now it remains a future plan. Instead, I used a so-called pseudo-distributed cluster, which is practically the development mode of Apache Hadoop.

The system consists of three main parts. The first part is a JavaScript module, which serves as a plugin to track the state of the devices used for browsing the monitored sites. This module communicates with the cluster which stores the data. The second part is the cluster itself, which is responsible for storing the collected data. The last module is a web application, which is responsible for displaying reports. I tried to make these reports as versatile as I could.

Downloads

Please sign in to download the files of this thesis.