OpenStack is a booming area of the storage ecosystem. From a handful of attendees a couple of years ago, more than 6,000 showed up at the recent OpenStack Summit in Vancouver. That said, OpenStack is something of a specialized area. A tiny fringe of the storage universe is enraptured by it while the rest wonders what the fuss is all about.
A useful analogy might be the early days of server virtualization. A small techie circle had uncontainable exhilaration about this new technology. Initially, the average IT guy said “Huh?” But that phase didn’t last long before virtualization became part of the mainstream. The same could well occur with OpenStack. But the first step is getting clued up on what it is.
According to Ross Turk, Director of Storage Product Marketing at Red Hat, OpenStack is the largest and fastest-growing open source project ever. He said the excitement around it is felt by all involved.
“The way infrastructure is deployed and consumed is changing as IT organizations are expected to cope with workloads that are increasingly dynamic,” said Turk. “OpenStack provides a platform for serving these needs: Resources that were once tightly-controlled and siloed are becoming available to end-users on-demand.”
Storage is a critical element for any OpenStack deployment. Instead of convenient and stable arrays and appliances, OpenStack requires storage that is flexible and elastic. Turk believes this allows enterprises and service providers to bring the flexibility and power of the cloud on-premise. By using OpenStack, IT departments can provide compute, storage and other resources to end-users in their organization on demand, he added.
How does OpenStack fit in with software defined storage (SDS)? Mike Karp, an analyst at Ptak Associates, said that OpenStack embraces SDS, but goes beyond it into what could be called a cloud operating system that controls large pools of compute, storage and networking resources throughout a data center.
“Everything is managed through a dashboard that allows users to provision storage through a web interface,” said Karp.
In a nutshell, then, OpenStack is an open source project that helps users deploy an infrastructure as a service platform. SDS specialist Nexenta notes that the adoption of OpenStack has increased tremendously over the last couple of years and has been moving from custom cloud environments to DevOps. OpenStack is now gradually finding its way into enterprise data centers as back-end infrastructures. With this transition, however, come new demands to support storage back-ends to be as scalable and elastic as the OpenStack infrastructure itself. Achieving this storage infrastructure of the future, said Oscar Wahlberg, Director of Product Management, Nexenta, requires scale-out architectures where users can grow capacity and performance quickly, easily and independent of each other.
“We are seeing OpenStack most commonly used in DevOps environments today, but moving closer to business-centric environments as the project matures,” said Wahlberg. “Today, we also see many large cloud and service providers use OpenStack to run their businesses and many large enterprise organizations are starting explore the benefits of OpenStack in private clouds.”
Greg Schulz, an analyst with StorageIO Group, puts things in perspective from the standpoint of traditional storage. He said that OpenStack is a collection of tools or bundles for building private, hybrid and public clouds. These various open source projects within the OpenStack umbrella include compute (Nova) and virtual machine images (Glance), dashboard management (Horizon), security and identify control (Keystone), network (Neutron), object storage (Swift), block storage (Cinder) and file-based storage (Manila) among others. It’s up to the user to decide which pieces you will implement. For example, you can use Swift without having virtual machines and vice versa.
“While OpenStack and traditional storage complement each other, you can continue to perform storage without OpenStack,” said Schulz.
OpenStack does require storage for its own internal needs, as well as space in which to host virtual machines along with their images.
“Similar to VMware and other software defined management stacks, some people will use OpenStack Cinders (block), Manila (file) or Swift (object) as one of their storage tiers,” said Schulz.
Others will delve more deeply into the OpenStack storage world while some might just dabble in it for now. Schulz thinks the service provider marketplace is where most of the adoption and deployment is occurring. There might be a few major enterprise projects ongoing, but it is far less publicized than what is happening with service providers.
“For some, OpenStack may well be the future for software defined storage management,” said Schulz. “But for others, it will be yet another storage tool or software.”
Vincent Hsu, CTO, IBM Storage, views OpenStack as an important element in what he calls the storage battleground in cloud environments. He said that its use of open standard APIs to optimize storage infrastructure for applications, analytics and data provides choice and total cost of ownership benefits. This enables users to move data seamlessly across devices, avoid vendor lock-in, scale performance and capacity, automate data management at scale and enable better collaboration, he said.
“OpenStack is critical to IBM's storage strategy in the era of cloud and software defined,” said Hsu. “The true value of SDS is when data is managed seamlessly across heterogeneous hardware and OpenStack allows that to happen.”
Photo courtesy of Shutterstock.