MATLAB distributed computing server
Mostrar comentarios más antiguos
Hello. Have a nice time. I've worked on data center clustering in the framework mapreduce.I wrote the program at MATLAB. And I want to use spark + hadoop with a quasi-distributive cluster. Does this need to install MATLAB distributed compute server?
Respuestas (1)
Kojiro Saito
el 9 de Abr. de 2019
0 votos
if you're using mapreduce and want to run in Hadoop, you don't need Spark.
There are two options.
1. Use MATLAB Parallel Server (MATLAB Distributed Computing Server until R2018b)
2. Use MATLAB Compiler to create mapreduce standalone application.
If you want to scale out tall array to Spark, there are two options.
1. Use MATLAB Parallel Server with Hadoop/Spark
2. Use MATLAB Compiler with Hadoop/Spark
Also, if you want to create stanalone applications wihch use MATLAB API for Spark,
1. Use MATLAB Compiler with Hadoop/Spark
6 comentarios
adele sharifi
el 9 de Abr. de 2019
Editada: adele sharifi
el 9 de Abr. de 2019
Kojiro Saito
el 9 de Abr. de 2019
Yes, in order to run mapreduce or tall in Hadoop/Spark, installation of MATLAB Parallel Server is necessesary even though just one PC.
adele sharifi
el 9 de Abr. de 2019
Kojiro Saito
el 9 de Abr. de 2019
adele sharifi
el 9 de Abr. de 2019
Kojiro Saito
el 10 de Abr. de 2019
Categorías
Más información sobre Third-Party Cluster Configuration en Centro de ayuda y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!