by Chris S
I have done my last 2 blogs on NoSQL databases and have introduced what they are and why they are becoming more and more popular. One particular type that was recently released by Neo Technologies was the Neo4j, which is a NoSQL graph type database. Graph data model, nodes with type relationships are joined across nodes. They attach a key value pairs to nodes and their relationships, relying more on pointers than indexes. What join allows is the ability to hop from one entity to another. Specifically a join is a CPU-bound operation that merges the criteria between the two tables. Therefore, by having a direct pointer instead of an index or merge, you end up with huge increases in performance. The Neo4j database is java based and has no SQL layer, but proves to be useful when wanting to take part of a database and use a graph database.
Eventually, I’m sure Neo will create a full on Graph Database that companies will be able to migrate data from a relational database over to the new graph database. NoSQL is showing its muscles when companies are searching for a database that provides scalability and performance. I look forward to seeing what other companies are planning to do with the NoSQL graph database type. So far Neo is the only company to release a project that utilizes this type, however other companies have released software using the other NoSQL types, such as MongoDB, and Apache Cassandra. Either way, this is another step in the right direction.
Kerner, Sean Michael, Neo Launches NoSQL Graph Database, October 25, 2011. Retrieved from databasejournal.com