Quantum Corp., provider of data protection and big data management, announced Thursday StorNext 4.3 software which brings new intelligence features, greater performance and increased scale to managing big data.
Incorporating a new database and supporting up to one billion files and dozens of petabytes of tiered storage, StorNext 4.3 pushes the boundaries for file system and archiving performance. This latest generation StorNext software also includes various capabilities for emerging customer needs such as archive on ingest, active vaulting and project-based capacity management -- all benefits designed to further manage the volume, velocity and variety found in massive unstructured data and to extract maximum value from it.
StorNext 4.3 is built to address emerging big data requirements in digital media content, genomics research, video surveillance and natural resource exploration.
Organizations are increasingly storing massive quantities of data to prepare for "long-tail" events, where only a small fraction of data will be required in the future, but predicting which data will become valuable is impossible. To store this data economically for the long-term, StorNext 4.3 supports Active Vault, a closely managed vault within the tape library that offers a much lower cost than the robot-accessible active shelves. User-defined, site specific policies automatically recommend tapes to be active vaulted, track their location, and monitor data integrity. When the data is needed again, an operator moves the tape from the Active Vault to the active shelves of the library.
When large numbers of files are ingested from sensors, cameras or other capture devices, the primary copy of data must reside on tape for economical reasons. Leveraging its new database architected on MySQL, StorNext 4.3 automatically protects big data and provides faster file truncation speeds, up to 1.5 times faster at 1 million files, to enable efficient "archive on ingest" at the scale of big data. When a file is written to a StorNext-managed repository's disk system, the software immediately archives or truncates the file to a large-scale tape library with minimal impact on ingest rates because only a small reference to the file remains stored on disk.
Compute intensive analytical applications, such as genetic sequencing and video transcoding, require high-speed file access, but not SAN speeds. StorNext's proprietary Distributed LAN Client (DLC) protocol leverages Ethernet to enable access speeds that are up to 50 percent faster than CIFS or NFS. With version 4.3, StorNext DLC Windows clients can experience a maximum throughput improvement of 360 percent over 10 GbE and 60 percent for 1 GbE.
Users who access and share big data content, such as editors and scientists, often work on projects that may be associated with multiple workgroups. The latest version of StorNext adds the flexibility and intelligence to manage storage capacity on a per-project basis through directory quotas. StorNext also improves archive performance for small files that are often archived along with large files, up to 40 percent faster writes for files that are less than 1 MB.
Many formats for storing big data use a series of multiple files versus a single very large file. To improve streaming read performance, StorNext 4.3 provides a new utility that can intelligently re-align files in these multi-file formats so they are stored contiguously on disk after retrieval from the archive. StorNext 4.3 also improves file creation times up to 50 percent, which improves overall write performance for multi-file formats such as DPX.
For Mac OS X only-based StorNext File System environments, StorNext 4.3 now supports Named Streams. StorNext 4.3 enables simpler hardware upgrades and disaster recovery. With the ability to migrate file system metadata stored on the metadata controller to a different disk geometry, the software allows customers to take advantage of new drive capacities and technologies.
StorNext 4.3 is available as a software-only solution or integrated in StorNext appliances. If purchased as software, Quantum now has a simplified pricing model for StorNext File System clients and a new slot-based StorNext Storage Manager offering, which applies to StorNext AEL Archives and third-party LTO library solutions. StorNext software and appliances can be purchased through Quantum's worldwide network of distributors and resellers, as well as from the company directly.
"When it comes to managing big data, there are several key challenges: huge volumes of data in a variety of formats, the velocity of record creation and variable latencies, and the complexities of individual data types within formats,” said Sheila Childs, managing vice president, Storage Technologies and Strategies, Gartner. “Organizations looking to adapt their infrastructures for big data would benefit greatly from selecting solutions which comprehensively address these issues."
"Quantum continues to push the limits and evolve StorNext with customers' big data management needs in mind. With re-architected tiered storage capabilities, StorNext enables customers to achieve greater scale, performance and reliability when managing content,” said Janae Stow Lee, senior vice president, Filesytem and Archive, Quantum. “This latest version of StorNext also leverages our vertical market expertise by including features for big data file formats and workflows that have specific requirements, thereby helping to maximize the value of the stored information."