GraphDB is used in scenarios where data is highly connected and the goal is to capture the complex relationships in the vastness of it. In a gist, it’s recommended where the Relationship between data is more important than the data itself.
Some of the well known Use Cases for GraphDB are
- Drive the customer engagement using the demographic data from previous similar and comparative engagements
- Obtaining 360-Degree Customer Viewpoint
- Recommendation Engine
- Used in Fraud Detection and Analytics to bring to light the unusual relationship between the entities
- Fault Analysis using anomaly detection
- Connect siloed data from CRM, Inventory, Tasks, Appointments and Point of Sales systems
- Privacy and risk compliance
Benefits of using GraphDB to support business requirements are due to the below facts
- Simplified Data Model
- Allow Flexible modeling
- Cater for both Structured and Unstructured Data
- Real-time Querying and update capability
- Simple Querying
- Parameterized Querying
- Graphic Visualization
- Test based evolution
GraphDB have the below mentioned Performance Advantages
- A basic query can run hundreds of times faster on a graph database than a traditional relational database.
- Graph-like queries, like finding the shortest path in a graph, are handled quickly and naturally.
- Graph databases do not rely on an index, so they can easily handle huge Volume of Data
- Graph databases do not rely on join operations.
- It can handle fluctuating data changes quickly due to the schema less architecture
Some Considerations while choosing a GraphDB for solving a business case
- Understand the graph use case before choosing a graphdb by defining business case, processing requirements and aligning with technology architecture
- The Query language complexity and ease of using ML capabilities
In the below table we can note that a GraphDB is better suited for querying data that is highly connected and has complex relationships. The Use-Case below represents a social network querying to find all the friends of a user’s friends. And furthermore for finding friends of friends of friends. Alexa and Jonas built this query in both MySQL and Neo4j with a database of 1,000,000 users and the results are striking. Execution Time is in seconds, for 1,000 users.
| Depth | Execution Time – MySQL | Execution Time –Neo4j |
| 2 | 0.016 | 0.010 |
| 3 | 30.267 | 0.168 |
| 4 | 1,543.505 | 1.359 |
| 5 | Not Finished in 1 Hour | 2.132 |
*Reference:Neo4j in action Jones partner and Alexa Vukoti
Below are some of the comparison points of the most popular implementations of graph databases – TigerGraph Vs Neo4j
| Name | Neo4J | TigerGraph |
| Data space on disk | No Data compression on Disk | Data Compressed on Disk |
| Query Language | Cypher Query Language | GSQL |
| Implementation Language | Java, Scala | C++ |
| Supporting Languages | .Net, Clojure, Elixir, Go, Groovy, Haskell, Java, JavaScript, Perl, PHP, Python, Ruby, and Scala | C++, Java |
| Multi Processing | APOC library, Cypher transaction Batching | Parallel computation built on BSP (Bulk Synchronous Parallel) |
| Operating System | Linux OS X Solaris Windows | Linux |
| License Model | OpenSource, Managed | Commercial, Managed |
| Deployment Platform | Cloud, On Premise | Cloud, On Premise |
| Supported Graph Algorithms | Neo4j Graph Data Science (GDS) | GSQL Graph Algorithm Library |
| Pricing Model | On Cloud: RAM, Consumption Based On Premise: machines/cores/RAM | On Cloud: CPUs and RAM size, pay-as-you-go hourly model. Free for 50 GB data, Annual Subscription: Tens of thousands of dollars for 50-100 GB and 1 million $ for 1 Terabyte of Data On Premise: Free for 50 GB data, Annual Subscription: Tens of thousands of dollars for 50-100 GB and 1 million $ for 1 Terabyte of Data |
– By Gayathri Bhave