Neo4j Data Importer & Workload Runner for LDBC Social Network Benchmark

This is a reference implementation of the LDBC Social Network benchmark, for the Neo4j graph database. It contains code for the following:

1. Importing datasets generated by the LDBC Social Network Data Generator into Neo4j
2. Running workloads from the LDBC Social Network Workload Definition against Neo4j

Build

Initial build:

https://github.com/ldbc/ldbc_socialnet_bm_neo4j.git
cd ldbc_socialnet_bm_neo4j
./build.sh

Subsequent builds:

mvn clean compile -Dmaven.compiler.source=1.7 -Dmaven.compiler.target=1.7

Import data into Neo4j

mvn exec:java -Dexec.mainClass=com.ldbc.socialnet.workload.neo4j.load.LdbcSocialNeworkNeo4jImporter

The resulting Neo4j instance will have this schema

Run Workload using ldbc_driver

usage: java -cp classesYouNeedOnTheClassPath.jar com.ldbc.driver.Client [-db <classname>] [-l | -t] [-oc <count>] [-P
       <file1:file2>] [-p <key=value>] [-rc <count>] [-s]  [-tc <count>] [-tu <unit>] [-w <classname>]
   -db,--database <classname>       classname of the DB to use (e.g. com.ldbc.driver.db.basic.BasicDb)
   -l,--load                        run the loading phase of the workload
   -oc,--operationcount <count>     number of operations to execute (default: 0)
   -P <file1:file2>                 load properties from file(s) - files will be loaded in the order provided
   -p <key=value>                   properties to be passed to DB and Workload - these will override
	                            properties loaded from files
   -rc,--recordcount <count>        number of records to create during load phase (default: 0)
   -s,--status                      show status during run
   -t,--transaction                 run the transactions phase of the workload
   -tc,--threadcount <count>        number of worker threads to execute with (default: 2)
   -tu,--timeunit <unit>            time unit to use when gathering metrics. default:MILLISECONDS,
	                            valid:[NANOSECONDS, MICROSECONDS, MILLISECONDS, SECONDS, MINUTES]
   -w,--workload <classname>        classname of the Workload to use (e.g.
	                            com.ldbc.driver.workloads.simple.SimpleWorkload)

Run Example Using Maven

MAVEN_OPTS="-server -XX:+UseConcMarkSweepGC -Xmx512m" 
mvn exec:java -Dexec.mainClass=com.ldbc.driver.Client 
-Dexec.arguments="-db,com.ldbc.socialnet.workload.neo4j.Neo4jDb,-w,com.ldbc.socialnet.workload.LdbcInteractiveWorkload,
-oc,10,-rc,-1,-tc,1,-s,-tu,MILLISECONDS,-p,neo4j.path=db/,-p,neo4j.dbtype=embedded"

The executed workload consists of these queries

Configuration

# Directory where CSV data can be found (from LDBC Data Generator)
data_dir=/path/to/ldbc/generator/social/network/csv/files/

# Directory to store Neo4j database instance
db_dir=db

# neo4j properties file for configuring the batch importer
neo4j_import_config=/neo4j_import_dev.properties

# neo4j properties file for configuring the transactional database
neo4j_run_config=/neo4j_run_dev.properties

Other

Using the LDBC Social Network data generator:

• Instructions on installing and running
• Explanation of format, content, and distributions, and schema of generated data TODO