Cirrus Data and the next frontier: Moving SAN blocks to the cloud

Cirrus Data believes that the new challenge for storage in the public cloud, up to now largely built on object and file data, is moving block data to the cloud, meaning mission-critical databases.

In the current economic environment, according to Ron Croce, Cirrus EVP for Sales and Market and CRO, enterprises are moderating their expenditure and halting planned storage refreshes. The company hopes, as others do, that customers will see moving data to the public cloud as a transfer of capital expenditure to operating expenditure.

Ron Croce

He told us: “The next frontier is block in the cloud.” 

With software like Pure’s Cloud Block Store the cloud destination can have a familiar on-prem SAN array interface. Vendors like Dell, with Project Alpine, and HPE are moving their block array software to the cloud. That provides a familiar SAN environment across the on-prem and cloud worlds for many of the installed SANs.

But the big obstacle is moving data there, and Cirrus Data can move block data to the cloud with no downtime and reasonably quickly. It can be seen as equivalent to a storage refresh in that there has to be a migration from the old SAN array to a new one, except that the new one is in a remote datacenter and running on storage instances rather than on a known storage array vendor’s hardware, albeit operating under the existing SAN vendor’s software abstraction layer.

According to Croce the biggest block data users are databases and the hardest to move to the cloud are high transaction rate DBMS – the mission-critical stuff that cannot have downtime. He said Oracle’s Golden Gate technology is a replication-based data protection product that can be used for migration, but claimed it could be disruptive, adding: “We do it five to seven times faster than a vendor’s own technology.”

He mentioned moving a 200TB Oracle database to the cloud for a large financial institution that took just a month from start to cutover. The customer couldn’t use ASM or Golden Gate for this.

Croce said: “All of the storage OEMs are partners of ours, except Huawei. … We are Pure’s exclusive partner for migration.”

He mentioned IBM, Infinidat, Hitachi Vantara and global system integrators like Kyndryl. Infinidat, he said, has only been using Cirrus’ appliance product so far but is about to adopt its Cirrus Migrate Cloud software product.

We can imagine Cirrus’ software being used for cloud-to-cloud data movement as well and, another interesting idea, since it moves blocks over time it could also do roll back and be used in data recovery scenarios.

We asked how Cirrus’ software can cope with SANs using NVMe-drives, which are not SANs in the traditional sense at all, as there are seemingly direct connections between the host application and the drives. He said: “That’s our future. We don’t see it being adopted yet. It’s five or so years away.”

The Cirrus engineering team is paying attention to this technology and Cirrus will support it when necessary.

Cirrus is also offering Migration-as-a-Service, with the software installed as part of a SAN and included as an operation during a storage refresh. This enables it to move on from being used just on a migration project basis and brings in recurring revenue.

If Croce is right, and moving block data to the cloud is the next frontier, then the future for block-based data movers looks bright. Certainly WANdisco, whose Live Datac tech can replicate Oracle and other data to the cloud, has reported a string of deals concerned with moving data to Azure and other clouds recently, and that supports Croce’s conviction that mission-critical data is moving to the cloud.