We use the term gigabyte (GB), as the binary (base 2) numeric scale for multiple bytes, and when we say GB we mean 1024 MB, as defined by the Joint Electronic Device Engineering Council (JEDEC). This unit of measurement is also known as a gibibyte (GiB).

How is the memory allocated?

Each Neo4j AuraDB instance is carefully calibrated to allocate the memory resources as a mix of JVM overhead, native memory and Java heap to be made available to the AuraDB Instance without you having to tune any of this. 

What does a CPU resource mean exactly and how can I compare this to a physical server single CPU performance?

Neo4j Aura runs on container technology and this allows for the service to allocate dedicated CPU resources to individual AuraDB Instances while allowing AuraDB Instances to burst shared CPU when needed. A CPU is in this case a vCPU or CPU thread.  

Why should you not compare AuraDB instances with a single physical server?

Comparisons of CPU in a physical server to the minimum dedicated CPU on Neo4j AuraDB is not relevant because it misses the real value of the service to our customers, including:   

AuraDB is Fully-Managed
Neo4j AuraDB takes care of security upgrades and Neo4j software patches automatically and invisibly without interruption of service. Further, our dedicated team of Neo4j Site Reliability Engineers monitor the service 24x7.

AuraDB data is safely backed-up
Neo4j AuraDB has automatic back-ups to cloud storage (preventing hardware failures) so your data is always copied and available for you to download or restore.

AuraDB is Always On
Neo4j AuraDB is always available. Whether it's invisibly applying a software patch or handling your request to scale your AuraDB Instance size up or down,. Neo4j AuraDB manages everything while maintaining Always On Availability.

AuraDB is Self-Healing
Should anything happen to the (sometimes fragile) infrastructure of the cloud, Neo4j AuraDB will find and repair the problem automatically before it affects your service. Neo4j AuraDB was built with high availability and data durability at its core.