TY - GEN
T1 - Distributed processing of elevation data by means of apache hadoop in a small cluster
AU - Komarkova, Jitka
AU - Spidlen, Jakub
AU - Bhattacharya, Devanjan
AU - Horak, Oldrich
PY - 2013
Y1 - 2013
N2 - Geoinformation technologies require fast processing of high and quickly increasing volumes of all types of spatial data. Parallel computational approach and distributed systems represent technologies which are able to provide required services, with reasonable costs. MapReduce is one example of such approach. It has been successfully implemented in large clusters in several instances. The applications include spatial and imagery data processing. The contribution deals with its implementation and operational performance using only a very small cluster (consisting of a few commodity personal computers) to process large-volume spatial data. Open-source implementation of MapReduce, named, Apache Hadoop, is used. The contribution is focused on a low-price solution and it deals with speed of processing and distribution of processed files. Authors run several experiments to evaluate the benefit of distributed data processing in a small-sized cluster and to find possible limitations. Size of processed files and number of processed values is used as the most important criteria for performance evaluation. Point elevation data were used during the experiments.
AB - Geoinformation technologies require fast processing of high and quickly increasing volumes of all types of spatial data. Parallel computational approach and distributed systems represent technologies which are able to provide required services, with reasonable costs. MapReduce is one example of such approach. It has been successfully implemented in large clusters in several instances. The applications include spatial and imagery data processing. The contribution deals with its implementation and operational performance using only a very small cluster (consisting of a few commodity personal computers) to process large-volume spatial data. Open-source implementation of MapReduce, named, Apache Hadoop, is used. The contribution is focused on a low-price solution and it deals with speed of processing and distribution of processed files. Authors run several experiments to evaluate the benefit of distributed data processing in a small-sized cluster and to find possible limitations. Size of processed files and number of processed values is used as the most important criteria for performance evaluation. Point elevation data were used during the experiments.
KW - Apache hadoop
KW - distributed processing
KW - elevation data
KW - small cluster
UR - http://www.scopus.com/inward/record.url?scp=84887091020&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84887091020
SN - 9789898565686
T3 - ICSOFT 2013 - Proceedings of the 8th International Joint Conference on Software Technologies
SP - 340
EP - 344
BT - ICSOFT 2013 - Proceedings of the 8th International Joint Conference on Software Technologies
T2 - 8th International Joint conference on Software Technologies, ICSOFT 2013
Y2 - 29 July 2013 through 31 July 2013
ER -