% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@INPROCEEDINGS{Grunzke:834093,
author = {Grunzke, Richard and Jug, Florian and Schuller, Bernd and
Jäkel, Rene and Myers, Gene and Nagel, Wolfgang E.},
title = {{S}eamless {HPC} {I}ntegration of {D}ata-{I}ntensive
{KNIME} {W}orkflows via {UNICORE}},
volume = {10104},
address = {Cham},
publisher = {Springer International Publishing},
reportid = {FZJ-2017-04094},
isbn = {978-3-319-58942-8 (print)},
series = {Lecture Notes in Computer Science},
pages = {480 - 491},
year = {2017},
comment = {Euro-Par 2016: Parallel Processing Workshops / Desprez,
Frederic (Editor) ; Cham : Springer International
Publishing, 2017, Chapter 39 ; ISSN: 0302-9743=1611-3349 ;
ISBN: 978-3-319-58942-8=978-3-319-58943-5 ;
doi:10.1007/978-3-319-58943-5},
booktitle = {Euro-Par 2016: Parallel Processing
Workshops / Desprez, Frederic (Editor)
; Cham : Springer International
Publishing, 2017, Chapter 39 ; ISSN:
0302-9743=1611-3349 ; ISBN:
978-3-319-58942-8=978-3-319-58943-5 ;
doi:10.1007/978-3-319-58943-5},
abstract = {Biological research is increasingly dependent on analyzing
vast amounts of microscopy datasets. Technologies such as
Fiji/ImageJ2 and KNIME support knowledge extraction from
biological data by providing a large set of configurable
algorithms and an intuitive pipeline creation and execution
interface. The increasing complexity of required analysis
pipelines and the growing amounts of data to be processed
nurture the desire to run existing pipelines on HPC (High
Performance Computing) systems. Here, we propose a solution
to this challenge by presenting a new HPC integration method
for KNIME (Konstanz Information Miner) using the UNICORE
middleware (Uniform Interface to Computing Resources) and
its automated data processing feature. We designed the
integration to be efficient in processing large data
workloads on the server side. On the client side it is
seamless and lightweight to only minimally increase the
complexity for the users. We describe our novel approach and
evaluate it using an image processing pipeline that could
previously not be executed on an HPC system. The evaluation
includes a performance study of the induced overhead of the
submission process and of the integrated image processing
pipeline based on a large amount of data. This demonstrates
how our solution enables scientists to transparently benefit
from vast HPC resources without the need to migrate existing
algorithms and pipelines.},
month = {Aug},
date = {2016-08-24},
organization = {European Conference on Parallel
Processing, Grenoble (France), 24 Aug
2016 - 26 Aug 2016},
cin = {JSC},
ddc = {004},
cid = {I:(DE-Juel1)JSC-20090406},
pnm = {512 - Data-Intensive Science and Federated Computing
(POF3-512)},
pid = {G:(DE-HGF)POF3-512},
typ = {PUB:(DE-HGF)8 / PUB:(DE-HGF)7},
UT = {WOS:000529303100039},
doi = {10.1007/978-3-319-58943-5_39},
url = {https://juser.fz-juelich.de/record/834093},
}