We are delighted to be a member of the Neo4j startup program!
This gave us the opportunity to upgrade our supply chain analytics graph database which is now a Neo4J cluster. For our customers this basically means
- improved reliability and scalability for our supply chain analytics and
- decreased analytics execution time.
We are now able to add additional hardware to the cluster at any time to keep up with the increasing number of analytics offered in the platform. At the same time our users will never suffer from the slightest downtime in supply chain analytics services, even if we encounter hardware problems in our Kubernetes cluster as the Neo4j architecture is now much more stable.
The Neo4j graph database always played an important role in sustainabill’s backend architecture as most use-cases in supply chain analytics reveal the weaknesses of traditional relational databases.
To query and aggregate data in a mid-sized supply chain for fast moving consumer goods, we need less than one second computation time with Neo4j where we would need about 7-8 seconds in our (actually really fast) relational database.
Why is that? Oversimplifying the truth a little bit, we basically need to find all relations from one facility or product to all other facilities or products involved in the upstream supply chain. These so called self-relations result in large join operations in traditional relational databases. Besides other improvements a graph database query uses nested loop joins instead of hash match joins which significantly decreases the join size and speeds up the queries.