Nada Chendeb Taher, Imane Mallat, Nazim Agoulmine, Nour Mawass
With the increasing number of connected things to the internet (IoT), the volume of the generated data and the rate at which it is generated are more and more increasing.In general, most of the generated data from these IoT devices is stored on some cloud infrastructure to insure scalability and continuous ease of access to it.With this situation, the concept of Big Data recently appears to lead the Artificial Intelligence and the Data Mining domains. In fact, if we use this ’Big Data’ correctly, we could turn it into ’Big Value’. Actually, the essential target of the emerging Big Data technologies is to provide techniques and tools to store large amount of complex data and prepare it to be analyzed and processed in order to get insights and predictions that could offer new opportunities towards better future.In this context, we have to deal with two main issues: the real-time analysis issue introduced by the increasing rate at which data is generated from IoT devices, and the long-term analysis issue introduced by the accumulation over time of huge volumes of data.In smart healthcare applications, these two issues appear clearly. In fact, medical sensors collect health related data from patients and send it to the cloud. This data should be analyzed permanently in real time to take appropriate decisions and act accordingly to save the patient’s life for example. The same data accumulated over time from different patients constitutes the bank of training data to build accurate machine learning model in order to perform smarter future disease prediction.In this paper, we propose an IoT-Cloud based solution for real-time and batch processing of Big Data. We use a Raspberry pi to replace the IoT device and generate data in real time. The objective is to provide a smart healthcare application into the cloud integrating IoT and Big Data techniques provided by the cloud operator to perform smart ECG analysis and early detection of any ECG anomaly. The solution was implemented and tested in AWS Amazon Cloud, it worked well and results show that the processing performance in term of response time for both long-term and real-time analysis is always guaranteed once the cloud resources are well provisioned.
Cloud Computing; IoT; Big Data Analytics; Healthcare; real-time analysis, batch processing