The evolution of data center storage architecture
A comprehensive collection of articles, videos and more, hand-picked by our editors
Startup Talena Inc. on Monday launched ActiveRx, a prescriptive analytics option for its Talena Big Data Availability Management platform.
Talena Big Data Availability software launched last year and can run on commodity x86 server hardware, inside a virtual machine or in the public cloud. The application is geared for secondary storage in big data environments, such as Hadoop, NoSQL databases and enterprise data warehouses.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Talena's big data platform uses a single copy of data to support multiple uses, including archiving and backup, development and testing, and disaster recovery.
Talena ActiveRx uses predictive intelligence on data availability and enables data analytics to be performed on passive backup.
The vendor said ActiveRx is expected to be available as part of an early-access program in April and generally available in late 2016. It will be sold under a separate software license from its core big data platform, which is sold as an annual subscription, priced per terabyte of storage.
Talena: Big data apps no longer a novelty
The analytics component is a response to the increased adoption of big data applications in enterprise storage, said Sanjay Sarathy, chief marketing officer at Talena. Predictive analytics seeks to find the best course of action among several choices.
"Big data applications used to be interesting research projects, but in the past five years, they have become tier-one applications for a lot of companies. Those applications are subject to [service-level agreements] and need always-on availability," Sarathy said.
"ActiveRx brings machine learning to bear on the workload. We can predict what your backup schedules are going to look like and when you might [miss] your recovery time objectives. We then automatically recommend any changes you need to make to the workflow," he said.
"You send your backup big data copy to Talena, and they back it up to software-defined, enterprise-class storage for point-in-time snapshots that users can mount directly," Matchett said. "It's not a full backup copy either; it's only as large as the deduplication ratio [allows]."
Using backup copies for live analytics
Talena Big Data Availability Management's agentless data-aware software gets installed on a primary node and introspectively recognizes when other big data nodes are added to the cluster. Its distributed file system pools direct-attached or networked storage capacity and uses application-specific APIs to move data.
Talena performs a full backup upon installation and incremental backups as schemas or tables get added to the cluster. Logical copies of data are stored on each node for high availability in the event of hardware failure.
Talena takes only one backup copy and uses proprietary block-level, post-processing deduplication to reduce the data footprint. Its file system supports multiple point-in-time snapshots for backup and erasure coding to survive multiple node failures. A distributed metadata catalog, known as Talena FastFind, lets users recover individual files or entire databases.
The vendor claimed it has several thousand nodes under management with customers in advertising technology, finance, media and entertainment, and retailing.
"Talena may be ahead of the inflection point, but the inflection point is inevitable," Matchett said. "The whole point of big data is being able to mash it together and see what you can find. Talena seems to be addressing the friction that occurs between big data and data protection."
Talena came out of stealth last August, with $12 million in funding. Its founder and CEO, Nitin Donde, was vice president of engineering for storage startups Intransa, Kazeon and Aster Data Systems. He briefly worked for EMC after it acquired e-discovery software startup Kazeon in 2009.
Take a big data storage tutorial
Robust storage systems required for big data analysts
Big data storage: The cloud vs. on premises