Leveraging stream processing engines in support of physiological data processing
Date
2018
Authors
Mishra, Sitakanta, author
Pallickara, Shrideep, advisor
Pallickara, Sangmi, committee member
Venkatachalam, Chandra, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Over the last decade, there has been an exponential growth in unbounded streaming data generated by sensing devices in different settings including the Internet-of-Things. Several frameworks have been developed to facilitate effective monitoring, processing, and analysis of the continuous flow of streams generated in such settings. Real-time data collected from patient monitoring systems, wearable devices etc. can take advantage of stream processing engines in distributed computing environments to provide better care and services to both individuals and medical practitioners. This thesis proposes a methodology for monitoring multiple users using stream data processing pipelines. We have designed data processing pipelines using the two dominant stream processing frameworks – Storm and Spark. We used the University of Queensland's Vital Sign Dataset in our assessments. Our assessments contrast these systems based on processing latencies, throughput, and also the number of concurrent users that can be supported in a given pipeline.
Description
Rights Access
Subject
predictive analytics
stream processing engines
real time data analysis
Internet of things