The new architecture disaggregates the storage from the compute, and is a key factor in both the much lower pricing and the greatly increased scalability.
Qumulo, which manages massive data stores, has released the next version of their Scale Anywhere Platform for unstructured data management. While this comes closely on the heels of the first version of Scale Anywhere, which came out in the spring, this version of the platform has been rearchitected to greatly improve performance and scalability, as well as make it more affordable.
“We originally launched Scale Anywhere in April,” said Ryan Farris, VP of Products at Qumulo. This is a major service evolution, not a relaunch. The other part of the announcement is Global Namespace [Q-GNS], a unified data plane for all of an organization’s unstructured data, that will go into private preview.”
The Scale Anywhere platform begins with Azure Native Qumulo [ANQ], a native cloud service which was jointly developed with Microsoft. The cloud-native enterprise filesystem provides unlimited, elastic scale, improved performance and breakthrough economics, to eliminate the need for customers to make trade-offs to get what they want.
Farris stressed how the pricing rehapes what has been standard in the market.
“$8 per TB per month has been the lowest cost available,’ he said. “We go as low as $30 per TB per month, for 1PB and beyond), with pricing starting at $37 per TB per month for a 100 TB commitment. This makes ANQ almost 80% less expensive than the closest alternative, and is comparable to the fully burdened on-premises cost of file storage.
“An architectural change which disaggregates compute and storage is how we can reduce our prices in Azure Blob this way,” Farris said. “It lets us scale elastically out and back in and only pay for burst periods. The previous version was more of a ‘lift and shift’ and this architecture is new as of this version.”
“One of the things we have to do as sellers is bring insights to our customers,” said Jeff Spears, Vice President of North American Sales at Qumulo. “This is a massive reduction from what people are used to seeing. VARs now can offer them a pricing program that gives them cloud for what they have been paying for on prem – so that they get 3x the performance.”
Ultimately, Spears emphasized, this gives customers the ability to make choices based on their business model and preferences rather than cost.
“The end customers are trying to sort out where they want to run workloads – on prem, or cloud,” he said. “This system gives them the flexibility to do both, and gives them more choice to run AI and ML on-prem or on the cloud.”
The flexibility, which scales to exabytes anywhere unstructured data is needed, also comes from the disaggregation of compute and storage.
“We don’t see another cloud storage service that’s truly elastic,” Farris said. “Our closest competitor in cloud is Azure NetApp files, but it doesn’t have our elasticity.”
In addition, ANQ gets 12 9’s of data durability from Azure ZRS blob, the best among services, can be configured and deployed directly from the Azure service portal in minutes, and provides a rich set of data services, including quotas, snapshots, multi-protocol access, enterprise security integrations, and real-time data analytics.
“It tiers to active or inactive archive, keeping cold data in cheaper storage and letting users operate on hot data in spoke sites,” Farris said. “Read caching will also be available in late January and early February.”