Installing Neo4j on Google Cloud with Compute Engine
Neo4j is a native graph database developed by Neo Technology, Inc. Native graph storage means that Neo4j stores data as nodes connected by relationships (there are no tables) while being ACID compliant. It also provides flexible and dynamic schemas that simplify (and speed up) queries.
Neo4j has high system requirements, especially in terms of memory. You should consider an instance having at least 4 GB of memory for a graph database of 0.5 GB and at least 8 GB for a graph database of 1 GB. If you need to run more than one Neo4j instance at the same time, given these high requirements, resource allocation may not be easy.
To solve that problem we’ve decided to use a cloud solution. The competition among cloud providers is pretty fierce at the moment, with Amazon, Google and Microsoft at the front. Despite the fact that we are Amazon users since ages, this time we opted for a High-Memory Compute Engine machine with local SSD from Google as its instances are roughly 50% cheaper than the similar ones from Amazon. Additionally, we created our own image with pre-installed and fine-tuned Neo4j to be used as a template for new machines.
Spinning new Google’s Compute Engine instances is easy and fast. We have convenient on-demand access to high-memory machines with the ability to control their usage i.e. there is no need to keep the instance running for the whole month, once the database computations are done, an instance can be discarded keeping the cost low. As a result we ended up with a flexible and cost-effective solution that perfectly matched our needs.
