Mobile Cloud Computing for Data-Intensive Applications

The computational and storage capabilities of today’s mobile devices are rapidly catching up with those of our traditional desktop computers and servers. In fact, mobile phones with 1 GHz processors are readily available in the market today. Unfortunately, all of these processing resources are mostly under-utilised and are generally used to process local data and programs only. With the use of local wireless networks, we can enable these phones to communicate with each other without utilising the resources of a global cellular network. This has the potential to enable collaborative data-intensive computing across a cloud of mobile devices without straining the bandwidth of global networks. To achieve these objectives, Hyrax [3] was initially developed by Marinelli as a MapReduce system [1] that is deployed on a networked collection of Android smartphones. Hyrax is based on Hadoop [4], which is a Java implementation of the MapReduce system and the Google File System [2]. While Marinelli has developed a system that is suitable for initial use to discover the resource constraints/challenges, performance and the scalability aspects of using mobile devices for collaborative data-intensive computing, that initial implementation of Hyrax was not suitable for wide-scale deployment on the mobile devices of common users. To that end, we have improved on Marinelli’s implementation of Hyrax, and aim to develop a mobile multimedia share-and-search application that allows users to discover relevant multimedia content on the mobile phones of those within reasonable proximity (ie within the same wifi network). We also evaluate the performance of this new implementation of Hyrax and identify areas for future work, especially in improving the performance and resource consumption (especially power resources) of Hyrax. Furthermore, we also consider the risk that the use of Hyrax poses to the users of the mobile devices that run Hyrax.