Pure Storage® launch of Pure Storage Cloud for Amazon Web Services (AWS) offers a way for Pure platforms to be extended into the cloud. In this blog/demo, see the benefits of running Cassandra on Pure Storage Cloud for AWS. I chose to demo Cassandra because it is one of the top 5 NoSQL databases in AWS based on this survey.
To demonstrate this, I built a 3-node cluster on Pure Storage Cloud, which is connected to another Pure Storage Cloud instance via ActiveCluster to provide maximum resiliency. The 3 node Cassandra cluster has data and commit logs on Pure Storage Cloud volumes.
I built another Cassandra cluster directly on EBS for comparison. AWS even has a best practices guide that recommends flash storage.
The following value propositions are covered in this demo:
- Data reduction of Cassandra data on Pure Storage Cloud
- Maximum Resiliency of Cassandra cluster on Cloud Block Store using ActiveCluster.
- Using Pure Storage Cloud instantaneous snapshots to protect or clone Cassandra cluster.
Data Reduction
The data reduction on Pure Storage Cloud is around 2.4 to 3.0 for the data and commit logs volumes. The data on Cassandra table was not compressed. This reduces the storage footprint by a big factor on Pure Storage Cloud. On AWS Elastic Block Store there is no data reduction.
Resiliency on Pure Storage Cloud
For maximum resiliency of the Cassandra on Pure Storage Cloud, it can be deployed on two instances of Pure Storage Cloud. These two Pure Storage Cloud instances are connected via ActiveCluster. The Cassandra data and log volumes on Pure Storage Cloud are added to ActiveCluster stretched POD. In the demo, to show resiliency one instance of Pure Storage Cloud is shutdown to simulate failure and then we check on the cluster status. As the instances are connected by ActiveCluster the data is synchronously replicated to the secondary Pure Storage Cloud instance. In case of failure of a Pure Storage Cloud instance, the Cassandra cluster will be up and running as if nothing happened. Pure Storage Cloud ActiveCluster here offers lot more than enterprise-grade resiliency.
Pure Storage Cloud snapshots
Pure Storage Cloud snapshots are instantaneous, which makes it a great feature to protect or clone your Cassandra clusters. In the demo I used them to recover the Cassandra cluster—also as mentioned before, it can be used to copy or clone the Cassandra cluster for development and testing purposes. It will be very difficult to do this using EBS snapshots. Cassandra offers native snapshots but it is very difficult to use. Most of the customers cannot take more than 2-3 native Cassandra snapshots before they run out of storage space, whereas Pure Storage Cloud snapshots are instantaneous and do not consume much space to start with. Also, we can take thousands of Pure Storage Cloud snapshots without any issues.
