Google launches storage services with knobs on

Google Cloud has expanded up its storage portfolio, augmenting existing services and launching a dedicated backup and data recovery service for the first time.

Google Cloud’s group product manager for storage, Sean Derrington, said the services were aimed as much at traditional enterprises as cloud native organizations, as both are looking to build resilient “continental scale” systems, and of course to drive down costs. A third aim was to support data rich applications, he said, which in “most cases” still feature a combination of on-prem as well as cloud data, meaning migration was always an issue.

So, in no particular order, Google Cloud Hyperdisk is described as a next generation complement to its Persistent Disk service block storage service, with different implementations for different workloads. Hyperdisk Extreme, for example, will support up to 300k IOPs and 4Gbps, to support demanding database workloads such as HANA.

But, Derrington continued, “We’re giving customers the option to basically have knobs that they can turn. Say within Hyperdisk Extreme, as an example, if I want to tune my IOPs to a certain level and I want my throughput to a lower level, because my application needs are different. And then I can also set the capacity.”

And if customers want to turn all the knobs to 11, they can, said Derrington, while they can also be adjusted over time as applications and workloads evolve.

The service will be rolled out in Q4.

For those who want to twiddle as few knobs as possible, Cloud Storage Autoclass will relieve storage admins of the drudgery of deciding what is hot and cold data and therefore where to keep it, i.e. Google’s standard, nearline, cold line, and archive cloud storage tiers.

Whatever tier data is in, it can be recovered in milliseconds, Derrington said. Google has also added Storage Insights to the Google storage console to highlight exactly what is going on within their data, for example, levels of duplication. This is in preview, with general availability in Q4.

As the birthplace of Kubernetes, it may not be a surprise that Google has launched a Backup for GKE service, offering backup and disaster recovery for K8s apps with persistent data. While there are third party services that offer this, Derrington said Google was the first hyperscaler. Again, the service will launch in early Q4.

It has also launched Filestore Enterprise multishare for GKE, which allows multiple pods – up to thousands – access to the same data, helping optimize storage utilization. It will be rolled out by the end of the year.

With the debut of a GKE backup service, it would seem odd for Google not to also launch a general Backup and Data Recovery service covering Google’s VMware Engine and Compute Engine platforms, as well as databases. Which is just what it has done.

“This is actually from the Actifio acquisition that we closed in December of 2020. This is now fully integrated into the Cloud Console,” said Derrington, and will be available this month.

Derrington said Google customers were perfectly free to continue using whatever backup and DR service they were using previously. “We do have an open ecosystem.”