Architecture

Prerequisites

• Install Node.js and npm
• Install Git
• Install and Configure AWS CLI
• The IAM user configured in the AWS profile must have the privileges to provision the following resources. Please note, while working with AWS acounts, IAM users and policies always follow IAM best practices:
  • Amazon VPC and Subnets
  • Amazon EC2 instance
  • Amazon S3 bucket
  • Amazon S3 Gateway VPC Endpoint
  • An Amazon Neptune cluster
• Install AWS CDK

Launching the app

To setup the app, run the following commands:

Download code from GitHub

git clone https://github.com/sahays/streaming-neo4j-msk-neptune
cd streaming-neo4j-msk-neptune/bootstrapper

Review configuration file

• Verify the app specific configuration saved in bootstrapper/cdk.json to ensure you have the right values for your environment
  • (Required) replace <your-key-pair-name> with your own EC2 key pair name e.g. my-us-west-2-key-pair
  • (Required) replace <provide-your-ip> with your current IP address e.g. 8.8.8.8/32 Know your IP address

{
  "app": "node bin/bootstrapper.js",
  "context": {
    "VPC_CIDR": "192.168.0.0/16",
    "EC2_CLASS": "t3a",
    "EC2_TYPE": "xlarge",
    "EC2_KEY_PAIR": "<your-key-pair-name>",
    "SG_FROM_IP": "<provide-your-ip>",
    "KAFKA_BROKER_INSTANCE_TYPE": "kafka.m5.large",
    "NEPTUNE_DB_INSTANCE_TYPE": "db.r5.large",
    "SOURCE_TOPIC_NODES": "Person{*};Movie{*}",
    "SOURCE_TOPIC_RELATIONSHIPS": "ACTED_IN{*}",
    "KAFKA_TOPIC": "graphstream"
  }
}

Deploy

After making changing to the file, run the following commands to launch:

npm install
npm run deploy

After successful run of the program (it takes about 10 minutes to complete), you'll see an output similar to the following:

Output	Purpose
.VpcID	Amazon VPC Id that contains all the resources
.MSKCluster	Amazon Managed Service for Kafka cluster ARN
.EC2Instance	Amazon EC2 instance Id
.NeptuneDbClusterIdentifier	Cluster name for the Amazon Neptune DB cluster
.NeptuneClusterEndpoint	Endpoint of the Amazon Neptune DB cluster

Detailed architecture

Post deploy steps

SSH into the Amazon EC2 instances created by the app. To see a list of all running containers run the following command:

docker container ls -a

You should be able to see the following 3 docker container services:

Service name	Purpose	Status
transformation-service	runs the transformation engine that transforms Neo4j data to Amazon Neptune data format	Running
neo4j-service	runs the Neo4j graph database version 3.5.6	Running
startup-service	runs the startup docker that fetches endpoint information from Amazon Neptune and Amazon MSK	Stopped
kafka-topic-service	creates a new topic in Amazon MSK	Stopped

If you want to see logs for a service, run the following command:

docker container logs <service-name>

Execute cypher scripts

To execute cypher scripts you need to enter the neo4j-service container using the following command

docker container exec -it neo4j-service cypher-shell

then execute a simple create command like the following

CREATE (TheMatrix:Movie {title:'The Matrix', released:1999, tagline:'Welcome to the Real World'});
CREATE (Keanu:Person {name:'Keanu Reeves', born:1964});
CREATE (Keanu)-[:ACTED_IN {roles:['Neo']}]->(TheMatrix);

Next, see the transformation-service logs by executing

docker container logs transformation-service

you should see an output similar to the following:

processing node {
  id: '0',
  before: null,
  after: {
    properties: {
      tagline: 'Welcome to the Real World',
      title: 'The Matrix',
      released: 1999
    },
    labels: [ 'Movie' ]
  },
  type: 'node'
}
inserted 0 {
  properties: {
    tagline: 'Welcome to the Real World',
    title: 'The Matrix',
    released: 1999
  },
  labels: [ 'Movie' ]
}
Movie
processing node {
  id: '20',
  before: null,
  after: {
    properties: { born: 1964, name: 'Keanu Reeves' },
    labels: [ 'Person' ]
  },
  type: 'node'
}
inserted 20 {
  properties: { born: 1964, name: 'Keanu Reeves' },
  labels: [ 'Person' ]
}
Person

Finally, to confirm that Amazon Neptune has been updated with streaming data run the following commands in order

docker run -it -e NEPTUNE_HOST --entrypoint /replace-host.sh sanjeets/neptune-gremlinc-345

:remote console

g.V().count()

Query architecture

The following code snippet shows the docker-compose environment section of the Neo4j service. This allows Neo4j to become a source of events for Kafka connector.

NEO4J_AUTH: none
NEO4J_dbms_logs_debug_level: DEBUG
# KAFKA related configuration
NEO4J_kafka_zookeeper_connect: ${ZOOKEEPER_CONNECT}
NEO4J_kafka_bootstrap_servers: ${BOOTSTRAP_SERVERS}
NEO4J_kafka_security_protocol: SSL
NEO4J_kafka_ssl_truststore_location: /var/lib/neo4j/temp/kafka.client.truststore.jks
NEO4J_kafka_acks: 1
NEO4J_kafka_num_partitions: 1
NEO4J_kafka_retries: 2
# streams
NEO4J_streams_procedures_enabled: "true"
NEO4J_streams_source_enabled: "true"
NEO4J_streams_source_topic_nodes_neo4j: Person{*};Movie{*}
NEO4J_streams_source_topic_relationships_neo4j: ACTED_IN{*}
NEO4J_streams_source_schema_polling_interval: 10000
# other
NEO4J_apoc_trigger_enabled: "true"
NEO4J_dbms_jvm_additional: -Djavax.net.debug=ssl:handshake
NEO4J_dbms_security_procedures_whitelist: apoc.*,streams.*
NEO4J_dbms_security_procedures_unrestricted: apoc.*,streams.*

The following code snippet shows the nodes and relationships that are configured to be the source of events for the Kafka connector. This ensures that whenever you perform CTUD operations on these nodes and relationships Neo4j will send changed data to Kafka. The transformation-service on the other had will recieve these data changes and will update Amazon Neptune

NEO4J_streams_source_topic_nodes_neo4j: Person{*};Movie{*}
NEO4J_streams_source_topic_relationships_neo4j: ACTED_IN{*}

The following architecture shows how neo4j-service works with transformation-service and Amazon Managed Service for Kafka.

Cleaning up

To cleanup AWS resources you need to run the following command:

npm run destroy