26 March 2016 Performance analysis of sliding window filtering of two dimensional signals based on stream data processing systems
Author Affiliations +
Proceedings Volume 9807, Optical Technologies for Telecommunications 2015; 98070Z (2016) https://doi.org/10.1117/12.2231384
Event: XIII International Scientific and Technical Conference on Optical Technologies in Telecommunications, 2015, Ufa, Russian Federation
Abstract
This research article contains an experiment with implementation of image filtering task in Apache Storm and IBM InfoSphere Streams stream data processing systems. The aim of presented research is to show that new technologies could be effectively used for sliding window filtering of image sequences. The analysis of execution was focused on two parameters: throughput and memory consumption. Profiling was performed on CentOS operating systems running on two virtual machines for each system. The experiment results showed that IBM InfoSphere Streams has about 1.5 to 13.5 times lower memory footprint than Apache Storm, but could be about 2.0 to 2.5 slower on a real hardware.
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Nikolay Kazanskiy, Nikolay Kazanskiy, Vladimir Protsenko, Vladimir Protsenko, Pavel Serafimovich, Pavel Serafimovich, "Performance analysis of sliding window filtering of two dimensional signals based on stream data processing systems", Proc. SPIE 9807, Optical Technologies for Telecommunications 2015, 98070Z (26 March 2016); doi: 10.1117/12.2231384; https://doi.org/10.1117/12.2231384
PROCEEDINGS
10 PAGES


SHARE
Back to Top