Parallel data intensive applications using MapReduce: a data mining case study in biomedical sciences

Performance is an open issue in data intensive applications (e.g. data mining tasks). Parallel and distributed computing systems (e.g. multicore computing, grid computing, cloud computing,etc.), along with hybrid programming models (e.g. MapReduce, MPI, etc.), is seen a sought-after solution for accelerating data-intensive applications. One of main challenges is how to exploit these advanced technologies effectively in facilitating fundamental science discoveries such as those in Biomedical Sciences. This paper explores how MapReduce and Cloud computing can accelerate performance of data intensive applications through a real data mining use case in the Biomedical Sciences. We have first adapted the data mining task using MapReduce model and then deployed it onto the Cloud. We have built an analytic model based on the MapReduce computations to evaluate the efficiency and performance of the prototype. The results, from both experiments and the evaluation model, show the performance and scalability can be enhanced through these advanced technologies.