Crunching Big Data with Julia
Tanmay K. Mohapatra
Introduce the big data infrastructure in Julia. It can be used to read/write HDFS files and run parallel Julia programs on a Yarn cluster.
This talk will use Elly.jl to demonstrate a big data workflow in Julia. Elly is a Hadoop HDFS and Yarn client. It is a pure Julia implementation with no dependencies on libhdfs. It provides:
- A familiar Julia ClusterManager interface, making it possible to use the familiar Julia parallel constructs on a Yarn cluster: addprocs, @parallel, spawn, pmap, etc.
- Lower level APIs to write native Yarn applications.
- A familiar Julia IO API for accessing HDFS files.
We shall use Elly and a few associated Julia packages to process a few example datasets.
Tanmay K.M., Julia contributor. https://github.com/tanmaykm