Red Hat to Acquire Gluster

October 4, 2011

By: Brian Stevens, CTO and VP, Worldwide Engineering, Red Hat

No one denies that cloud is disruptive. Taken together with the explosion of unstructured data, whose growth is dramatically outpacing that of structured data, enterprises are challenged to design new strategies for managing both compute and data. Legacy solutions are highly challenged in solving this emergent need.

Today we announce the acquisition of Gluster, Inc., and by doing so extend the Red Hat solution portfolio with industry-leading capability for managing the storage of unstructured data. On the cloud, on-premise, or bridged between the two.

Since the introduction of Red Hat Enterprise Linux almost ten years ago, we have enabled our customers to migrate to robust, enterprise-grade application platforms deployed on commodity x86-class servers. Delivering increased performance at lower cost. Developed by the largest engineering team in the world -- the open source community -- resulting in a roadmap of continuous refinement, optimization and innovation. Today many of the most demanding large-scale enterprise workloads are deployed on open source and commodity hardware.

Much of the build-out of IT starting in the 1990s and continuing into the past decade centered around automation of businesses processes. For these workloads, the structured data captured in traditional relational databases and SAN storage served as the workhorse, and these technologies continue to fulfill an important role today. However, this past decade has also witnessed the widespread shift from analog to digital formats that, when combined with the near-universal build-out of broadband network connectivity, has fueled massive demand for new classes of storage. Storage for capturing pictures, audio, video, medical images and documents. Storage to support mobile devices, for social media data, for capturing rich personalized customer interactions. While estimates vary, unstructured data is believed to account for over 80% of all data, and continues to grow at a rate that significantly outpaces structured data.

This is a segment where dramatic reductions in storage costs make whole new classes of workloads economically viable. Much of the early innovation in this space was built on an open source foundation, occurring in places like Google, Facebook, YouTube, Yahoo and others. Emerging from this work is the open source-based SMAQ (Storage, Map/Reduce and Query) stack, analogous to the open source LAMP stack that is so pervasive for web-centric applications. Underpinning this approach is massive scale-out, sharding and replication of data and the ability to exploit inexpensive commodity hardware.

As this trend towards digitization, mobility and connectedness continues, the modern enterprise is increasingly facing similar challenges, both in terms of new opportunities and threats to the status quo. For the many enterprises currently optimized around traditional databases and SAN storage solutions, this unstructured data explosion risks both breaking the wallet and hitting the scalability wall. A different approach is called for.

With our experience in bringing the benefits of open source to the enterprise, Red Hat began researching open source-based storage solutions that will address the issues of tomorrow:

  • Ability to massively scale out in a modular, incremental manner based on evolving demands
  • Optimized for storage and management of unstructured data
  • Architected for cost-effectiveness based on Linux and commodity hardware technologies (x86, Gigabit networking, and SATA/SAS drives) without the need for custom hardware
  • Enabling a choice of hardware platform and vendor
  • Able to deliver consistent response time globally via efficient replication
  • Supporting industry standards including POSIX, NFS and CIFS
  • Leveraging existing scalable, high-performance Linux filesystems
  • Deployable both on-premise, and on private and public clouds
  • Ability to securely replicate data into the cloud as applications require

An impossibly high bar, but one the team working on GlusterFS has made the center of their design since the project began six years ago.

Gluster and its leadership team began with a mission of advancing enterprise capability for reliably storing and high-performance processing of unstructured data. Since the core architecture of GlusterFS is a software only scale-out file system, it is unique in its ability to provide a single solution for on-premise, private, public and hybrid clouds.

 

One of the challenges of scale-out filesystems is management of metadata: the information of where specific files are located within the complex. Many distributed filesystems employ a centralized metadata repository, which both creates a potential scalability bottleneck and common point of failure or corruption. Gluster employs a unique no-metadata server model, based on an elastic hashing algorithm. This is a highly scalable and robust approach, which has no theoretical scalability limits and, as an example, is already employed in clusters with over 500 servers and petabytes of data.

Today, Gluster technology, when combined with Linux, can be used to address a number of important use-cases:

  • A fully distributed, highly scalable filesystem in support of data-intensive Linux applications , to provide big-data analytics
  • Exposed via NFS and CIFS, providing NAS services to Linux, UNIX and Windows servers and desktops
  • Integrated with KVM and other virtualization technologies, providing a scalable filesystem for storing virtual machines and for private and public clouds, as well as the ability to deploy NAS services at the tenant level for existing public clouds
  • An object model accessed via REST APIs for client and web-centric workloads

IT organizations can easily deploy an on-premise storage solution, an all-cloud deployment or a unification of the two into a hybrid cloud.

At Red Hat we have long recognized that open source enables the world's most advanced development paradigm. It comes as no surprise that GlusterFS was created by a worldwide collaboration of developers and end users. Over the past six years, a vibrant community has formed around GlusterFS, led by Anand Babu (AB) Periasamy, Gluster CTO. Red Hat has participated directly within the Gluster community, working on areas of multi-tenant security and new replication models. Last month the Gluster technology received the InfoWorld 2011 Best of Open Source (BOSSIE) Award, in addition to previous recognition from Storage Magazine, Citrix Synergy and TiE50. This year they were named as one of the top 25 coolest vendors by CRN.

But value is to be measured by the end user. And today users from Pandora, Box.net, Deutche Bank, Samsung, Autodesk and others depend on Gluster technology to address demanding problems in the fields of Education, Energy, Finance, Government, Media, Communications, Retail, Sciences, Medicine, BioTech and Technology Services.

For those of you within the Gluster community today, we thank you for your participation in making Gluster into what it has become today. In return we offer our continued commitment to its ongoing open development.

And for Gluster customers, we offer the assurance that the experience you depend on today to run your business will now be backed up by the innovative team at Gluster, and by Red Hat, the worldwide leader in open source solutions.

Back to top