kovaleff - Fotolia

MapR Technologies launches enterprise file storage product

MapR expands beyond analytics into the enterprise storage market with an exabyte-scale data store designed to manage files and containers in data centers and across clouds.

Hadoop distributor MapR Technologies Inc. is expanding into the enterprise storage market with software designed to manage files and containers in data centers and across the internet of things and clouds.

The MapR-XD exabyte-scale data store is an extension of the San Jose, Calif., vendor's MapR Converged Data Platform. The MapR platform also includes analytics and machine learning engines, a NoSQL database and event data streaming capabilities.

MapR Technologies has been a major contributor to Apache Hadoop open source projects. Its MapR Converged Data Platform integrates the company's intellectual property and open source code, such as Hadoop, Spark and Apache Drill.

The new MapR-XD software runs on commodity hardware and is based on the distributed file system that MapR Technologies has sold to customers for approximately five years. The file system did not have its own product SKU until MapR-XD hit the market last week.

Enhanced for flash, tiering

With the product release, MapR enhanced the file-system software with flash I/O optimization, REST API support and global namespace multi-tiering capabilities. MapR-XD will let users move data among hot (flash), warm (disk) and cold (cloud) options.

MapR's global namespace means applications don't need to worry about where the data is, said Bill Peterson, senior director of industry solutions at MapR. "If an application is using data that has been aged out to the cloud, the application doesn't need to know that," Peterson said. "It still just runs the query."

MapR-XD features include replication, point-in-time snapshots, distributed mirroring and a management console to enable administrators to check the health of their storage environments. The product supports a range of APIs, including NFS, POSIX, Amazon Simple Storage Service and Hadoop Distributed File System, as well as Amazon Web Services, CenturyLink, Google and Microsoft Azure public clouds.

The MapR-XD software also includes a client to give containers access to persistent data. The company added the MapR Persistent Application Client Container earlier this year in response to customer requests for technology that can support stateful applications, Peterson said.

Separate flash and disk licenses

MapR Technologies sells separate MapR-XD licenses for flash and disk.

Customers must purchase a flash license to move data from flash to disk or from disk to flash. The product supports standard SAS and SATA SSDs and emerging nonvolatile memory express-based PCI Express drives.

Peterson cited the new flash and tiering capabilities, as well as the number of customers added in the last year, as reasons the company decided to market its MapR file system.

MapR focuses on Fortune 1000 companies and has about 800 customers, with approximately half using the software on premises and the other half in the cloud, Peterson said. Its customer base spans vertical industries, such as telecommunications, finance, healthcare, automotive and manufacturing, Peterson added. Users often need to combine structured and unstructured data to be able to run analytics on it, he added.

Because we have the global namespace, you can have all of that data in one cluster and make it available for analytics or artificial intelligence.
Bill Petersonsenior director of industry solutions, MapR Technologies

"Because we have the global namespace, you can have all of that data in one cluster and make it available for analytics or artificial intelligence," Peterson said.

SAP uses MapR-XD for the storage layer of its cloud platform, and Samsung runs it for the information repository of its Galaxy phone's Bixby virtual assistant, Peterson said.

MapR-XD runs on Linux and supports a range of commodity hardware options from Cisco, Dell, Hewlett Packard Enterprise and Lenovo. MapR declined to disclose the starting list price other than to say it uses a subscription model based on capacity under management.

Big data roots

MapR Technologies made its name in the big data market, said George Crump, founder and president of Storage Switzerland LLC. "Big data's a good area to earn your stripes and then move into broader storage offerings," he said. "But it's still a file system, so it's not going to go after the SAN or iSCSI market. This is very similar to what we're seeing a lot of people do using data lakes -- a central collecting point that performs well and can be accessed by a variety of different file protocols."

MapR's competition spans scale-out NAS products, such as Dell EMC's Isilon; software-defined storage startups, such as Elastifile and Hedvig; and any general-purpose, highly scalable file system, including IBM's Spectrum Scale, Crump said.

"What's unique about [MapR] is where it came from," he said. "It came from a market that is all about scalability and distributed systems, and it's sort of backfilling the expected stuff. It has a skill set in an area that most legacy storage vendors would kill to have, so it'll be interesting to see what happens. The challenge it's going to have to overcome is its own success to a degree, in that it was almost too well-known in big data."

Henry Baltazar, a storage research director at 451 Research, said MapR will benefit organizations where line-of-business people have greater say in storage decisions. Many would prefer a product integrated with their analytics, he said, and it's unlikely an enterprise would use MapR-XD without a need for analytics.

MapR's competition would also include analytics players, such as Hortonworks, as the analytics specialists try to crack the data storage market, and storage vendors that have added analytics capabilities to their software, Baltazar said.

"Pretty much every IT player right now has a storage solution they're already using for file storage," said Scott Sinclair, an analyst at Enterprise Strategy Group Inc. in Milford, Mass. "Those solutions are likely not as optimized for app dev and analytics as MapR is. So, [MapR] will be absolutely competing against the inertia of, 'Can I use what I have already, and would it be good enough?'"

John Webster, an analyst at Evaluator Group Inc. in Boulder, Colo., wrote via an email that he expects MapR-XD to have broad appeal across enterprises and cloud service providers. But, he added, "They are competing in a storage world where they are not well-known."

Next Steps

Learn more about MapR Hadoop distribution

MapR's Hadoop adds Streams message system

Big data platforms pose challenges for new users

Dig Deeper on Storage management and analytics

Disaster Recovery
Data Backup
Data Center
and ESG