AWS EC2 R5 instance types on the Instaclustr Managed Platform
Continuing our efforts in adding support for the latest generation AWS EC2 instances, Instaclustr is pleased to announce support for the EC2 R5 instance type for Apache Cassandra clusters on the Instaclustr Managed Platform. The R-series instances have been optimized for memory intensive applications that require high-performance databases, distributed in-memory caches, in-memory databases, and big data analytics. On the Instaclustr Managed Platform, they are suitable for Cassandra and Kafka clusters when they are required by customer applications to deliver very high performance I/O with minimal latency.
The latest generation of R-series (R5s) instances provide significantly better price-to-performance and price-per-GB over its predecessor. They are designed to gain the most out of the hardware capabilities allowing applications to attain the best value for money. R5s also have EBS-optimized burst support, delivering better peak throughput than R4s. Read more about AWS EC2 R5 instances here.
If you are interested in migrating your existing Cassandra and Kafka cluster on the Instaclustr Managed Platform from R4s to R5s, our highly experienced Technical Operations team can provide all the assistance you need. We have built several tried and tested node replacement strategies to provide zero-downtime, non-disruptive migrations. Read our Advanced Node Replacement blog for more details on one such strategy for Cassandra.
Apache Kafka Benchmarking Results
Our Kafka performance benchmarking methodology aims to find the maximum consumer throughput (messages consumed per second) for a given test set. For this test a large number of producer load generators were established, enough to stress our Kafka cluster. Under this peak load, a consumer group was also established to consume messages from Kafka, essentially achieving the maximum consumer throughput. The results showed 16% increase in consumer throughput with R5s compared to R4s.
The load test set was comprised of two 3-node Kafka clusters, one with R4.xlarge and the other with R5.xlarge instances. Each node had 750GB EBS backed SSDs attached. A 30 partition 3x replicated topic was created on both the clusters. The performance was calculated on small 100 byte messages. Average message consumption per second was 1,994,499 messages/sec with R4s, and 2,322,540 messages/sec with R5s.
We are working on benchmarking R5s for Cassandra and will soon publish the results once the testing is completed.
If you want to know more about this benchmarking or need clarification on when to use the R5 instance type for Cassandra of Kafka, reach out to our Support team (if you are an existing customer), or contact our Sales team. You can also refer to our support article that describes the nature of each offered instance type and their likely use cases.
R5 series is slightly cheaper than R4s and the cost savings are passed to our customers. But considering R5’s improved performance, the overall price-performance gain is significantly higher. You can access pricing details through the Instaclustr Console when you login, or contact our Support or Sales teams.