Better Together: Delivering Graph Value with AWS & Neo4j - Antony Prasad Thevaraj
0 likes118 views
This document discusses deploying Neo4j, a graph database platform, on Amazon Web Services (AWS). It provides information on the partnership between Neo4j and AWS, how Neo4j fits into the AWS ecosystem, best practices for architecture and configuration, and steps for deployment using the AWS Marketplace or Neo4j Partners GitHub repository. The document aims to help users get started with running Neo4j on AWS infrastructure.
Better Together: Delivering Graph Value with AWS & Neo4j - Antony Prasad Thevaraj
- 1. © 2023 Neo4j, Inc. All rights reserved. © 2023 Neo4j, Inc. All rights reserved. 1 N E W Y O R K
- 2. © 2023, Amazon Web Services, Inc. or its Affiliates. Neo4j on AWS The Leading Graph Data Platform Antony Prasad Thevaraj Senior Partner Solutions Architect, AWS
- 3. © 2023, Amazon Web Services, Inc. or its Affiliates. Neo4j and AWS Partnership ● >40% of Neo4j customers run on AWS ● Member of AWS Partner Network since 2013 ● Collaborative Joint Engineering ● APN Advanced Tier Partner ● Data and Analytics Competency ● SageMaker Service Ready ● AWS Marketplace Seller ● AWS Public Sector Partner ● AWS ISV Workload Migration ● APN Global Startup ● ISV Accelerate
- 4. © 2023, Amazon Web Services, Inc. or its Affiliates. Real-Time Recommendations Fraud Detection Network & IT Operations Master Data Management Identity & Access Management Risk & Compliance REVEALING NEW WAYS TO SOLVE THE WORLD’S MOST PRESSING BUSINESS PROBLEMS
- 5. © 2023, Amazon Web Services, Inc. or its Affiliates. Neo4j in the AWS Ecosystem AWS Cloud Connector for Apache Kafka Connector for Apache Spark Amazon SageMaker Amazon Managed Streaming for Apache Kafka Amazon EMR Neo4j Enterprise Edition Graph Data Science Graph Database Bloom Data Warehouse Connector Amazon Redshift
- 6. © 2023, Amazon Web Services, Inc. or its Affiliates. Neo4j and SageMaker - Knowledge Graphs for Domain Specific LLMs Create a knowledge graph in Neo4j. Use an LLM such as Hugging Face on SageMaker to query it and build conversational systems 1. Amazon Textract and Comprehend process varied data sources 2. Entities are written to a Neo4j knowledge graph with relationships modeled. 3. Users can then query a SageMaker LLM model powered by Hugging Face 4. That converts natural language queries into Cypher graph query language AWS Cloud AWS CloudFormation Stack Amazon SageMaker Neo4j EE Graph Database Amazon Comprehend Amazon Textract User Queries Apps Structured Data Unstructured Data Knowledge Discovery Business Insights Graph Data Science
- 7. © 2023, Amazon Web Services, Inc. or its Affiliates. Neo4j and SageMaker - Graph Feature Engineering Engineer novel features in Neo4j. Use them to improve the accuracy of supervised learning models built in SageMaker. 1. Import data into Neo4j Graph Database 2. Generate feature embeddings in Neo4j Graph Data Science 3. Export the embedding to SageMaker Autopilot 4. Train and deploy an Autopilot model AWS Cloud AWS SageMaker Domain AWS CloudFormation Stack Amazon S3 SageMaker Default Bucket AWS SageMaker Studio Neo4j EE AWS SageMaker Autopilot Python GDS API SageMaker API SageMaker API Graph Database Graph Data Science
- 8. © 2023, Amazon Web Services, Inc. or its Affiliates. Best Practices
- 9. © 2023, Amazon Web Services, Inc. or its Affiliates. Architecture Bloom Public subnet Public subnet VPC AWS Cloud Availability Zone 1 Availability Zone 2 NAT gateway NAT gateway Neo4j Leader Neo4j Follower Internet gateway Public subnet Availability Zone 3 Neo4j Follower Auto Scaling group NAT gateway Neo4j Disk Storage Business Analysts Data Scientists Developer Developer/DBA Client Access BOLT: 7687 HTTP: 7474 HTTPS: 7473 DRIVERS Application Browser BOLT: 7687 HTTP: 7474 HTTPS: 7473 BOLT: 7687 Neo4j Disk Storage Neo4j Disk Storage
- 10. © 2023, Amazon Web Services, Inc. or its Affiliates. Recommendations & Considerations Compute Starting Out: - t3.medium or similar First Data Workload: - r6i.large Production Workload: - r6i.xlarge Storage Starting Out: - 50GB Min. (EBS) First Data Workload: - 100GB Min (EBS) Production Workload: - 100GB Min (EBS) Networking Connectivity: - Neo4j access allowed from 0.0.0.0/0 by default - SSH access can be specified using a parameter in CloudFormation
- 11. © 2023, Amazon Web Services, Inc. or its Affiliates. Deployment
- 12. © 2023, Amazon Web Services, Inc. or its Affiliates. Neo4j on AWS Marketplace ● Neo4j Enterprise ○ AMI and CFT ○ BYOL ○ Graph Database ○ Graph Data Science ○ Bloom ● AuraDB Enterprise ○ DBaaS ○ Annual Commitment ○ Private Offers
- 13. © 2023, Amazon Web Services, Inc. or its Affiliates. Neo4j Partners GitHub ● Deploy with the AWS CLI ● Source code for the IaaS Marketplace listing ● Simple template that is a good starting point for customizations
- 14. © 2023, Amazon Web Services, Inc. or its Affiliates. Next Steps ● Try it yourself! ○ AWS Marketplace ○ Neo4j-Partners GitHub ○ AWS Quick Start ● Further Reading ○ Landing Page ○ APN Partner Finder