Red Hat this week announced that it's shipping a new version of its Gluster cluster file system, which the company has renamed Red Hat Storage Server 2.0. Red Hat also unveiled a partner ecosystem for the product that includes Cisco, Groupware, Intel, Mainline and Synnex. Initially released as a beta version April 9, Storage Server is the industry's first file storage system to integrate easily with object storage, according to Ranga Rangachari, VP and GM of Red Hat's Storage Business Unit.
Since April, the product has been through a beta program with 40 large-enterprise customers worldwide, in which it has been used to address new storage projects, says Rangachari. "We have a lot of customers across multiple verticals, with 90% of their storage growth unstructured."
Rangachari says in addition to general-purpose file server applications, customer uses can be broken down into three broad groups: cloud-based content creators looking for seamless growth; the expanding need for near-line storage; and high-performance computing. Open source storage systems and volume x86 servers have the potential to transform the storage market in the same way that Linux and volume x86 servers transformed the server market, and Red Hat is positioning itself to be at the forefront of the industry transformation, he says.
The company first launched the Red Hat Virtual Storage Appliance for Amazon Web Services (AWS) in February, shortly after it acquired Gluster. Since its acquisition in October 2011, Gluster has been working on integrating its storage products into the other Red Hat units, such as JBoss and the cloud business unit.
Based on the Red Hat Enterprise Linux operating system, Storage Server 2.0 also features compatibility with more than 50 dual-socket x86-based servers; support for Common Internet File System, Network File System, HTTP and OpenStack Swift protocols; and geo-replication for data protection and disaster recovery. Currently available as only a technology preview but due out around the fall is support for Apache Hadoop and a management console based on the oVirt Project, an open source infrastructure and virtualization management platform that aims to provide visibility and control into storage clusters via a single pane of glass, says Rangachari.
In her review of the beta release, Deni Connor, a principal analyst at Storage Strategies Now, said instead of using a metadata server, Red Hat Storage Server 2.0 uses a hashing algorithm to locate data in the storage pool, which allows it to remove I/O bottlenecks and single points of failure. It's also Red Hat Enterprise Virtualization-ready and includes improved management capabilities, such as support for network lock manager and improved self-healing.
According to a prepared statement from Terri McClure, a senior storage analyst at Enterprise Strategy Group, Red Hat is helping to pave the way for a fundamental change in how storage is sold and consumed, with software-based, scale-out unified file and object storage, deployable on-premises or in the public cloud, supporting unstructured and semi-structured big data. Gartner's Stan Zaffos, VP and research director, storage, says users are desperate to contain storage costs, and scale-out storage built using open source storage software running on x86 servers configured with direct-attached storage components offer an attractive low cost option to classic storage arrays.
Looking ahead, Rangachari says, there will be a focus on tighter integration with virtualization during the next few quarters. "That will really start to set us apart, either as a storage or virtualization vendor."