Lightweight multi-language bindings for Apache Spark

Luca Salucci, Daniele Bonetta, Walter Binder

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

Apache Spark has emerged as one of the most prominent frameworks for distributed high-performance data analysis. Among Spark’s most appealing features are its bindings for dynamic languages such as Python and R. Despite of the great flexibility of such languages, they often cannot match the performance of statically typed languages such as Java or Scala. However, this limitation is not only due to the intrinsic nature of dynamically typed languages. Largely, the performance gap is caused by the way the language runtimes interact with Spark. In this paper we describe a new approach to integrating Python and R into data-intensive Spark applications. Our approach significantly reduces the performance gap between such languages and their statically typed counterpart, making dynamic languages an attractive alternative for the implementation of big-data applications.
Original languageEnglish
Title of host publicationParallel Processing - 22nd International Conference on Parallel and Distributed Computing, Euro-Par 2016, Proceedings
EditorsP.-F. Dutot, D. Trystram
PublisherSpringer Verlag
Pages281-292
ISBN (Print)9783319436586
DOIs
Publication statusPublished - 2016
Externally publishedYes
Event22nd International Conference on Parallel and Distributed Computing, Euro-Par 2016 - Grenoble, France
Duration: 24 Aug 201626 Aug 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on Parallel and Distributed Computing, Euro-Par 2016
Country/TerritoryFrance
CityGrenoble
Period24/08/1626/08/16

Fingerprint

Dive into the research topics of 'Lightweight multi-language bindings for Apache Spark'. Together they form a unique fingerprint.

Cite this