Skip to content

giga-dev/insightedge

 
 

Repository files navigation

InsightEdge

Documentation: User Guide
Community: Slack Channel, StackOverflow tag, Email
Contributing: Contribution Guide
Issue Tracker: Jira
License: Apache 2.0

InsightEdge is a Spark distribution on top of in-memory Data Grid. A single platform for analytical and transactional workloads.

Features

  • Exposes Data Grid as Spark RDDs
  • Saves Spark RDDs to Data Grid
  • Full DataFrames API support with persistence
  • Geospatial API for RDD and DataFrames. Geospatial indexes.
  • Transparent integration with SparkContext using Scala implicits
  • Data Grid side filtering with ability apply indexes
  • Running SQL queries in Spark over Data Grid
  • Data locality between Spark and Data Grid nodes
  • Storing MLlib models in Data Grid
  • Continuously saving Spark Streaming computation to Data Grid
  • Off-Heap persistence
  • Interactive Web Notebook
  • Python support

Building InsightEdge

InsightEdge is built using Apache Maven.

First, compile and install InsightEdge Core libraries:

# without unit tests
mvn clean install -DskipTests=true

# with unit tests
mvn clean install

To build InsightEdge zip distribution you need the following binary dependencies:

Package InsightEdge distribution:

mvn clean package -P package-community -DskipTests=true -Ddist.spark=<path to spark.tgz> -Ddist.xap=<path to xap.zip> -Ddist.zeppelin=<path to zeppelin.tar.gz> -Ddist.examples=<path to examples.zip>

The archive is generated under insightedge-packager/target/community directory. The archive content is under insightedge-packager/target/contents-community.

To run integration tests refer to the wiki page

Quick Start

Build the project and start InsightEdge demo mode with

cd insightedge-packager/target/contents-community
./sbin/insightedge.sh --mode demo

It starts Zeppelin at http://127.0.0.1:8090 with InsightEdge tutorial and example notebooks you can play with. The full documentation is available at website.

About

InsightEdge Core

Resources

License

Contributing

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages

  • Scala 65.0%
  • Shell 22.5%
  • Batchfile 6.5%
  • Java 5.2%
  • Python 0.8%