Today we announced Cloud Block Store, our new cloud data service on AWS which is an industrial-strength cloud block storage offering that enables mission-critical applications like Oracle to run in the cloud. The strength of this service is that it extends the Pure FlashArray™ capabilities into the cloud which facilitates use cases that were arduous to envision earlier.

With Cloud strategy still evolving and as IT is moving towards Hybrid Cloud, many Organizations are considering placing their Oracle DR in the cloud.  Now with the introduction of Cloud Block Store, this use case is very easy and simple to build.

There are two ways the Oracle DR can be set up on the Cloud.

  1. Using Oracle data guard to set up a physical standby database on the Cloud
  2. Using native storage replication to set up asynchronous replication between on-premises and the Cloud.

With Cloud Block Store, either of these options can be done with ease but also provides the following additional benefits on the Cloud:

  • Efficient block storage:  Experience the benefits of deduplication, compression, and thin provisioning on the Cloud that deliver capacity and performance
  • Instance space saving Snapshots: Use Cloud Block Store to instantly create clones of the physical standby database for development, quality testing, reporting, etc.,
  • Always on Encryption:  All data that goes into the Cloud Block Store is encrypted to protect your asset on the Cloud
  • Active/Active Metro Clustering: The Cloud Block Store can be deployed across multiple availability zones with Purity ActiveCluster to gain high availability

Note: It is not a requirement to have Pure FlashArray on-prem to use the Cloud Block Store.  If you are already a Pure customer you are very well aware of all the value propositions of Pure FlashArray and we are extending that value into the Cloud.  

Following two use cases were performed to show the simplicity and ease of use of Cloud Block Store on AWS.  

1. Instant database cloning of the Physical Standby on Cloud Block Store

The environment includes an Oracle Linux 7.5 database server running an Oracle 12.2 production database (PROD) on Pure FlashArray within our datacenter.  Native snapshot of the PROD database was replicated to the Cloud Block Store on AWS to set up the initial physical standby database (PROD_SBY) hosted on an EC2 instance (oradb02) running Linux 7.5 and the standby was kept in sync using the Oracle redo streams. (Please see the link to set up physical standby database using Pure FlashArray snapshots)

While our test environment included Pure FlashArray on-prem, you could have your production database on any storage on-prem.  In such case, you have to follow the regular RMAN backup, transfer the backup to AWS to create the physical standby database on the Cloud Block Store.  

To showcase the instant database cloning of the database on AWS, we performed the following steps three times to create DEV, QA and STG databases.  All these databases were co-hosted on another EC2 instance (orawsdb) on AWS.

  1. A record was inserted into the PROD database on-prem.
  2. Performed a redo switch to sync the data into the standby database.
  3. Stop the managed recovery on the standby database.
  4. Take a native snapshot of the protection group that comprises the standby database volumes on Cloud Block Store.
  5. Start the managed recovery on the standby database.
  6. Instantiate the protection group snapshot to new volumes on Cloud Block Store.
  7. Attach the volumes to the EC2 instance through iSCSI and discover the volumes using
  8. Create the cloned database out of the new volumes within Oracle context.

The entire process was automated through shell scripting and the steps 4 and 6 are performed on the Cloud Block Store using the command line interface.

While the source standby database allocated space is 67GB and used space is 27GB, the actual space usage of the DEV, QA, and STG volumes are significantly lower as they were provisioned out of space-efficient snapshots and will start consuming space only when changes are performed on these databases.


2. Data Mobility of an Oracle database to Cloud using native replication

Pure’s Cloud Block Store enables seamless data mobility between on-prem and the cloud using the array native replication functionality.  The asynchronous replication uses the space efficient snapshot functionality to replicate the data between on-prem and the cloud. The asynchronous replication is very efficient that it only sends the data that is deduped and compressed across the wire.

To perform the data mobility of an Oracle database between on-premises and the cloud using Cloud Block Store we inserted a record on to the source database (OLTPROD), took a native snapshot of the source database volumes, replicated it to the Cloud Block Store, instantiated the snapshot on the Cloud Block Store to a new volume and created a new database named OLTPQA.  

The following screenshot shows the script that inserts a record and takes a snapshot of the two volumes which are then replicated to the Cloud Block Store.

In a short span (within seconds), the Cloud Block Store shows the status of the replication of the protection group as completed.

The following screenshot shows the cloning process on the EC2 instance on AWS that instantiates the snapshot by copying them into volumes that were mounted through iSCSI on the EC2 instance, creating the controlfile for OLTPQA and open the database in resetlogs mode.

The following screenshot shows the end of the demo which took only 26 seconds to complete and also shows the record that was inserted prior to the snapshot at the source.

Here is the demo video that shows both the use cases on Cloud Block Store in detailed steps.

Using Pure’s Cloud Block Store, we just scratched the surface with a couple of simple and powerful use cases.  Pure’s Cloud Block Store uncovers numerous use cases for disaster recovery, high availability, data mobility, continuous integration, the continuous development between on-premises and the cloud.  

We are just getting started on this exciting journey of unifying cloud and this blog just covered one of the three cloud data services that we announced today.  Read more about the other cloud data services and Pure’s cloud vision in our announcement blog.