These days, it’s hard to read an IT periodical without some reference to cloud computing, and it’s hard to read about cloud computing without seeing some reference to OpenStack. Next week kicks off another OpenStack Summit, this time in Tokyo, at which thousands of developers, architects, system administrators, and business professionals reconvene to discuss and share what’s new with one of the fastest growing and most popular open source technologies since Linux. It’s the true power of the open source community on center stage. And while the impact of OpenStack on IaaS and cloud computing infrastructure is hard to deny, the impact of software-defined storage on OpenStack is becoming the same way.
By Toshinori baba (Own work) [CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons
Think about OpenStack as a Japanese bullet train racing down a freshly laid railway, like investments in cloud computing. These investments are expected to grow at a 30% CAGR from 2013 through 2018, compared with 5% for overall enterprise IT. (as reported in "Cloud State of the Union, 2015, citing "Riding the Cloud Computing Wave; RHT Down to Sell," Goldman Sachs, 13 Jan. 2015.) And by 2016, it is projected that about 80% of enterprises worldwide will be using IaaS. ("The Role of Cloud in IT Modernisation, the DevOps Challenge," Ovum, 18 November, 2014). The only problem is catching the OpenStack train gracefully, and that’s where software-defined storage enters the picture.
THE RIGHT KIND OF STORAGE IS CRITICAL TO OPENSTACK
Let’s face it. Businesses know that if they don’t embrace the cloud, they risk falling behind. But if they deploy incorrectly, they can jeopardize the data that’s core to their lifeline. They’re going to miss the train and might even derail it. They need a ticket, and software-defined storage is their ticket to ride! In fact, OpenStack is inextricably linked to storage — like Fred Astaire and Ginger Rogers, or a web application and its database, and inadequate storage can lead to cloud infrastructures with severely limited capacity and usefulness. Traditional NAS and SAN architectures are often too rigid in scope and in design. Storage appliances can be high performing, but they’re generally proprietary, monolithic, hardware-based complexities not well-suited for petabyte scale, resulting in isolated data silos that aren’t able to adapt to a next-generation OpenStack workload. This is why savvy OpenStack users are embracing software-defined storage. It’s their ticket to board the train before it speeds away without them.
CEPH IS PREFERRED
The most popular storage used by OpenStack operators and application developers is, hands-down, Ceph. (OpenStack User Survey, May 2015.) Red Hat Ceph Storage has joined Red Hat Gluster Storage to constitute the Red Hat Storage portfolio, a collection of software-defined storage championed by Red Hat in concert with the open source community. Why Ceph? Because Ceph provides unified infrastructure capabilities to support block, object, and file storage under one roof; because it’s tightly integrated with OpenStack’s modular architecture and key components for ephemeral and persistent storage including Nova (compute), Cinder (block), Glance (images), Swift (object), and, via GlusterFS, Manila (file); because it’s massively scalable to manage petabytes of data; and because it’s flexibly configurable to adjust storage needs as applications and deployment change — exactly the type of dynamic behavior one expects with OpenStack.
Something else people don’t realize. More than a decade ago, the founders of Ceph technology set out to engineer an optimal, massively-scalable, software-defined storage system for cloud infrastructure using a flexible scale-out architecture. OpenStack was released in 2010 with the same general concept of providing foundational IaaS technology. Ceph’s architecture (including the block storage device) therefore predates OpenStack and was actually the inspiration for its Cinder block storage abstraction layer. This says a lot about the logic and suitability behind the two technologies working together, including enterprise-grade solutions like Red Hat Ceph Storage combined with Red Hat Enterprise Linux OpenStack Platform or Red Hat Cloud Infrastructure.
So the next time you read about OpenStack or prepare to deploy it, remember its tie between OpenStack and Ceph. Ceph is your ticket to the midnight train to OpenStack. And if you happen to be in Tokyo, stop by and see us in the Red Hat booth (P7) or attend one of our Ceph Collaboration Day sessions on Wednesday afternoon. Can’t make it to the Summit this time? Then check us out at redhat.com/storage or follow us throughout the Summit @redhatstorage.