Announcing R5d instances and lookup cache for Amazon Neptune

Starting today, you can launch your RDF/SPARQL or Apache TinkerPop graph application using R5d instances on Amazon Neptune. The R5d instance type is based on the Amazon EC2 Nitro System and includes local NVMe-based SSD block level storage. Neptune R5d instances introduce a lookup cache that leverages the low-latency NVMe SSD storage to improve read query performance and reduce data retrievals from storage.

For use cases with frequent, repetitive lookups of property values or RDF literals, R5d instances use the lookup cache to temporarily store these values in the NVMe SSD volume for faster retrieval in the future. R5d instances can be deployed in the same Neptune cluster as R5 instances, enabling customers tailor the workload to the use case and customizing the compute instance needed for specific types of graph workloads. When the lookup cache is fully warmed with repeated access of property values or RDF literals, Neptune R5d instances can also observe a decrease in the amount of data read from storage and achieve better memory utilization.

R5d instances are supported with Amazon Neptune database running engine version 1.0.4.2 and higher. The lookup cache is only available on R5d instances, where it is automatically enabled by default.

To learn more about the lookup cache and check if your use case could benefit, refer to the Amazon Neptune User Guide. For more information on pricing and region availability, refer to the Neptune pricing page and AWS Region Table.