DataGravity Inc. today launched its first product, the Discovery unified hybrid array series that includes search, data governance and data protection features.
The Discovery Series 2000's integrated discovery and search engine is designed to allow IT teams, security, compliance and business users to automatically analyze information from unstructured files. It provides data governance through real-time user activity tracking capability, along with data protection for real-time file restores.
DataGravity was co-founded by Paula Long -- also a co-founder of iSCSI storage pioneer EqualLogic -- and John Joseph, former EqualLogic marketing vice president. Dell acquired EqualLogic in 2008. Long and Joseph have raised $42 million in funding for their new company.
The first two Discovery products are the DG2200 (48 TB) and DG2400 (96 TB), which are identical except for their storage capacities. Discovery combines hard disk drives with flash-optimized cache, and support for CIFS, NFS, SMB and iSCSI storage protocols. The primary controller node stores data on hard drives, and mirrors the data to a secondary intelligence node that mines the data. Moving data to the intelligence node prevents production I/O from being impacted.
Discovery's software crawls through primary storage metadata and data, showing users the correlation of data across people, time and content. Discovery also has a user interface with authenticated, role-based access designed for multiple users.
"We watch activities and data interaction," said Joseph, who is DataGravity's president. "Today, storage systems are dark. We enable users to see data within the storage system. The DataGravity engine is attached to the array and opens the data. We look at the content of over 400 different file types. We look at who owns the file, whether it was changed and track the changes. It's like a Facebook timeline for a file."
The search and discovery engine allows users outside of the IT department to view the history of a file. It tracks user activity, even within virtual machines in a VMware environment, user access profiles and has the ability to identify who accessed the content. The array does file-level restores, along with dynamic protection policies and storage allocation.
"We know when a file is corrupted, and our indexing function can find the file when it was not corrupted, so you can decide which file to recover," Joseph said. "I can create a file share of 100 GB and decide I want to protect it by selecting the snapshot frequency based on time, percentage of data changes or the gigabyte amount of data changed."
DataGravity CEO Long describes Discovery as "a new template" for storage that is designed to allow business owners to gain the most value from their data. List price for the arrays ranges from $50,000 to $100,000, including all software features.
"This is not Hadoop in a box," she said. "This is a way of looking at your data, mining it for what it can tell you, and presenting it in a homogenous way so you can ask questions of the data and solve business problems."
Auditing and search functions key for beta/alpha users
Beta tester Chris Berube, IT manager for the Sanford, Maine, law offices of Joe Bornstein, said he tested the array by copying data from his production environment to the DataGravity 92 TB configuration. He said one of its key functions is its ability to look at data inside a data store without having to log in to VMware. He also cited the auditing capability as one of the important functions it delivers.
"What I care about is who accessed the file and who touched it," he said. "Who accessed, who touched it, when [it] was read, what changes were made and what it look[ed] like before the change. It gives you basic file, security and auditing. It's almost like e-discovery for my type of [firm]. The word e-discovery just gives me nightmares. This is a good in-between."
Mark Lamson, director of technology at Westerly Public Schools in Rhode Island, was a DataGravity alpha tester. He said he found its search capability useful for Microsoft SQL databases.
"We have a lot of databases that run on SQL, but the front-end application does not give you good tools to search," he said. "You have to have a lot of in-house expertise and you have to cobble something together. [With DataGravity], we found personal identifiable information and that is something you want to protect. We found a payroll director file. We were able to check if anyone else touched that file. In traditional storage, it would be like looking for a needle in a haystack."
DataGravity drops clues about storage management product
Deciding between a hybrid or all-flash array