Neo4j the benefits of graph databases + a NoSQL overview

Another form of NoSQL databases that we have not discussed in the greatest of lengths are graph databases. Well that is going to change, in this post we are going to discuss Neo4j a graph database written in Java. The presentation, by Emil Eifrem CEO of Neo Technology, at 80 slides it's packed with a ton of information.

Some key highlights from the presentation

Four trends are helping to drive NoSQL

1. Data Set Size
2. Connectivity
3. Semi-Structure
4. Architecture

Key points about graph databases in general

• Inspired by Euler and Graph Theory
• Core abstractions are nodes, relationships between nodes and properties on both
• Example implementations, AllegroGraph and Neo4j

Key system characteristics of Neo4j

• Disk-based
• Transactional (unique characteristic when compared to other NoSQL datastores)
• Scales up (billions of nodes on single VM)
• Robust (6+ years in production)

Other interesting facts

• Multiple query languages available: SPARQL (SQL for linked data) and Gremlin (perl for graphs)
• Multiple language bindings: Jython, CPython, JRuby, Clojure and Scala

Anyway, there is so much information in the slides I really can't do it enough justice with just bullet points, check it out!

New player in the NoSQL movement, OrientDB

Interesting new participant in the NoSQL space, OrientDB. What is OrientDB?

OrientDB is a new Open Source NoSQL DBMS born with the best features of all the others. It's written in Java and it's amazing fast: can store up to 200,000 records in 5 seconds on common hardware. Even if it's Document based database the relationships are managed as in Graph Databases with direct connections among records. You can traverse entire or part of trees and graphs of records in few milliseconds. Supports schema-less, schema-full and schema-mixed modes. Has a strong security profiling system based on user and roles and support the SQL between the query languages. Thanks to the SQL layer it's straightforward to use it for people skilled in Relational world.

In addition to the features above it supports transaction. A feature that is often missing from other NoSQL data stores. I’ve added links both to the code hosting site and the project site as well. One other piece of info is that OrientDB is the creation of one individual, which is both impressive and scary at the same time. While it is released with the Apache license and code is available it does not yet have a huge community supporting it.

Orient Home Page
Orient Project Page (Google Code)

Graph Databases, NOSQL and Neo4j

So far a majority of the posts here on NoSQLDatabases.com have been about a number of different types of NoSQL data stores such as the column, key/value or document based solutions. One area we have not focused on is graph oriented data stores. Well that officially changes now.

InfoQ presents a very in depth post by Peter Neubauer, who happens to be one of the founders of the open source graph data store Neo4J. In the post he points out in his mind two very important aspects of NoSQL data stores, scalability and complexity.

With regards to scalability the article has this to say:

Many of the NOSQL databases above all have loosened up the requirements on Consistency in order to achieve better Availability and Partitioning. This resulted in systems know as BASE (Basically Available, Soft-state, Eventually consistent). These have no transactions in the classical sense and introduce constraints on the data model to enable better partition schemes (like the Dynamo system etc). 

Peter goes on to say the following about complexity of the data that we are trying to store in relational databases:

The increasing interconnectivity of data as well as systems has led to denser data sets that cannot be scaled and autosharded in any obvious, simple or domain-independent way, even noted by Todd Hoff.

The post goes into greater detail about the relational model vs. the object graph and gives several examples of code in both Java and Ruby. It’s a really good post and recommend you check it out.

Read More: Graph Databases, NOSQL and Neo4j