Web Browser-based Social Distributed Computing Platform applied to Image Analysis
Autores: Mikel Zorrilla, Ángel Martín, Iñigo Tamayo, Naiara Aginako, Igor G. Olaizola
Fecha: 02.10.2013
Abstract
In this paper we introduce a new platform to perform image processing algorithms over big data. The main stakeholders of media analysis are the social services which manage huge volumes of multimedia data. While social service providers have already a big resources pool of connected assets through the devices of the community, they are not exploiting them for their processing needs and they usually deploy high performance systems that run batch works. Image processing requires parallelizable atomic and lightweight tasks that can benefit from a big community of thin devices executing seamless background processes while the user enjoys other social media contents. To provide such infrastructure a client-side browser solution based on JavaScript libraries has been developed. We also describe a performance model that establishes the contexts where the solution gets ahead in terms of available resources and the processing problem nature.
BIB_text
author = {Mikel Zorrilla, Ángel Martín, Iñigo Tamayo, Naiara Aginako, Igor G. Olaizola},
title = {Web Browser-based Social Distributed Computing Platform applied to Image Analysis},
abstract = {
In this paper we introduce a new platform to perform image processing algorithms over big data. The main stakeholders of media analysis are the social services which manage huge volumes of multimedia data. While social service providers have already a big resources pool of connected assets through the devices of the community, they are not exploiting them for their processing needs and they usually deploy high performance systems that run batch works. Image processing requires parallelizable atomic and lightweight tasks that can benefit from a big community of thin devices executing seamless background processes while the user enjoys other social media contents. To provide such infrastructure a client-side browser solution based on JavaScript libraries has been developed. We also describe a performance model that establishes the contexts where the solution gets ahead in terms of available resources and the processing problem nature.
}
isi = {1},
date = {2013-10-02},
year = {2013},
}