Advanced Computing in the Age of AI | Thursday, August 5, 2021

Quantum’s New All-Terrain File System Brings Data Insights to Applications 

Data storage vendor Quantum’s new All-Terrain File System (ATFS) platform will launch in December, giving enterprises the ability to integrate real-time data classification and insights with critical business applications on network-attached storage systems.

The all-new, file-based ATFS platform, which is designed to make it more efficient for enterprises to view their data while automating purposeful data placement based on policies, was recently unveiled by Quantum as part of several related feature enhancements within its StorNext 7 File System and ActiveScale object storage products.

By integrating real-time data classification and insights more closely with applications, Quantum ATFS is built to help enterprise users determine how their storage resources are allocated and consumed, according to the company. This information can be used by enterprises to have more direct oversight of its data repositories, rather than guessing about their sizes and capacities. Quantum ATFS is built to help enterprises bring together data from a wide range of siloes to make it more usable and easier to access in serving the business. ATFS can be used to control and monitor data on-premises or in the cloud.

The ATFS file system can work with data that is stored in flash, bulk, or on the cloud using policies, application defined tags, or manual configurations, according to Quantum. The file storage platform can be tuned based on the size of the active data set, while automated policies can place data just-in-time to support workloads more efficiently.

Quantum ATFS control portal

ATFS can be used to automate application workflows by integrating with asset management tools, schedulers, and other applications to automate tasks using API in life sciences, media and entertainment, finance, and more. It also includes active data retention tools, such as metadata and tags that simplify access to data over time, as well as the ability to deploy resources in the cloud using cloud-based applications. Quantum ATFS also includes enterprise collaboration features that allow secure data sharing across the organization and externally without creating duplicates. A wide range of tools to control and secure data are also built in, serving requirements including data retention, protection, regulated access guidelines, best practices and more.

“To leverage data to advance AI and machine learning, you need to know what data exists, who owns it, and which rules are associated with it,” Eric Bassier, the senior director of marketing for Quantum, told EnterpriseAI. “ATFS aligns an application’s needs with storage resources in real time and provides deeper insights to data. This is a huge step toward enriching that data for the purposes of research.”

Quantum ATFS was born to help with these kinds of problems, he said. “Unstructured data sprawl has become a common refrain for organizations that depend on an intake of video, digital imagery, sensor data, and any form of unstructured data. We have customers with millions or even billions of files and no real visibility into what they have or where it resides. When you combine this visibility with rapid data growth, the pressure on IT and storage infrastructure can be enormous.”

The ATFS services will become available to order in December as software installed on a Quantum appliance.

ActiveScale Gets New Features

Quantum’s ActiveScale 5.7 object storage platform now includes an Object Lock feature that prevents data from being deleted or moved during malicious attacks, including ransomware incidents. Using the object lock feature, immutability is set for an object or a bucket, making it non-changeable until the set policy expires, which guards it from attacks. The object lock can also be used for compliance or protection against data loss, ensuring that the data remains in its current state for as long as desired by the user.

Also new to the ActiveScale object storage line is an entry-level model in a smaller, three-node configuration starting at 432TB raw. The smaller version allows customers with smaller data storage needs to get started with ActiveScale at a lower price but with all the management, protection, and preservation capabilities that they need, according to the company.

The ActiveScale platform also gets one other update –  a new Small Object Aggregation capability that aggregates small files into a large object prior to erasure encoding it, which enhances its ability to provide efficient storage capacity. The aggregation results in higher utilization and increased overall performance for small object transactions, according to Quantum. The key differentiator of small object aggregation is the ability of ActiveScale to provide direct access to small objects without reconstituting the whole object.

StorNext 7.0 Updates

Also getting updates is the latest version 7.0 of Quantum’s StorNext file system, which now includes a new File System Pools feature that automates data placement on NVMe and HDD storage for high-throughput, low-latency workloads. The StorNext File System Pools enable administrators to define pools of NVMe, SSD, and HDD within their file system and create policies to move files and folders between those pools automatically. This gives them the benefit of files that move seamlessly between pools for greater performance at a lower overall cost, according to Quantum.

StorNext 7.0 also gets an easier to use, redesigned and more intuitive user interface aimed at helping enterprises better classify, manage and protect their unstructured data. StorNext 7 includes expanded web services APIs that provide new ways to query metadata, automate data movement, configure and manage the file system. It also comes with a simplified, capacity-based licensing model to make it easier for users to monitor their costs and usage. The new version will be generally available by mid-December.

Analysts on Quantum’s Updates

Randy Kerns, a senior analyst and strategist with Evaluator Group, said the latest Quantum storage announcements will be helpful for enterprise users. “ATFS is somewhat revolutionary in that it has integrated data management into the NAS system, beyond what you would see in a typical NAS,” said Kerns. “ActiveScale and StorNext are logical evolutionary improvements.”

The ATFS debut gives enterprises the ability to gain a greater view of their data and automation by using the catalog of information created when data is ingested, said Kerns.

The newly-unveiled, entry-level ActiveScale model will be an economical storage repository for AI and ML data for enterprises, said Kerns, allowing them to start smaller at lower cost and have the flexibility to grow in the future. “The StorNext changes will give more control to users for data placement and therefore performance. We see ATFS more in the enterprise datacenter space. It’s good to see Quantum continuing to invest in development.”

Another analyst, Steven Hill, a senior research analyst for applied infrastructure and storage technologies at 451 Research, said Quantum’s latest news reflects a growing industry-wide focus on the challenges of managing, securing and insuring the availability of unstructured data.

“This is more or less the natural evolution of business data itself, in that more and more business applications spanning nearly every vertical market are generating and utilizing [more and more] file-based data,” said Hill. “The quantity of unstructured data in the mix is starting to exceed that of traditional database information in enterprise storage, as well as for data that’s to be used for long-term analytics and AI/ML applications. The unstructured data challenge is at zettabyte scale now, and will only get worse as initiatives like IoT have the potential to dramatically increase the amount of file-based data.”

That’s where products and services like Quantum’s latest offerings come in, he added.

“Unstructured data presents much greater management challenges, in that most file-based information offers little or no insight into its contents or provenance; which makes in extremely difficult to establish appropriate management policies, much less provide insight into its business context and value for analytics purposes,” said Hill. “A growing number of storage vendors are now looking to provide a new model for handling unstructured data that leverages the capabilities of next-generation, hybrid storage systems to play a larger role in granular storage management; offering governance based on information the data holds rather than what it was randomly named or where it physically resides.”

Add a Comment

EnterpriseAI