Intelligent Infrastructure

Mass data and next-gen workloads Intelligent Infrastructure

Enterprises slowly but steadily adopt open source storage software

Open source storage software has  gotten off to a slow start in the enterprise

The commercial viability of OpenStack, Hadoop and associated open source cloud storage technologies is still hotly debated by enterprises and service providers. While many prominent vendors, from Cisco to IBM and many of Seagate's Cloud Builder Alliance partners, have made substantial contributions to community-driven projects, customers have been more measured in their uptake of these technologies to support critical cloud-based workflows.

Some of the advantages and drawbacks of open source
Open source initiatives usually have the advantage of being cutting-edge, while the software that emerges from them is often much more economical than proprietary alternatives. The widespread success of Linux-based operating systems, MySQL and Apache Lucene demonstrate how these efforts can be both fast-moving and relevant to the enterprise.

Still, doubts remain about open source, especially in regard to its feasibility for tasks such as enterprise cloud storage. For example, many industry watchers have noted the fundamental do-it-yourself nature of OpenStack, which has resulted in a surplus of unique deployments but an absence of a reference architecture that would likely speed up adoption among enterprises.

Accordingly, commercial solutions based on OpenStack in particular, although integral to that project's long-term viability, have been slow to take off. Some organizations have instead turned to public cloud offerings such as Amazon Web Services that offer versatile ecosystems and APIs in addition to global scalability.

Buyers that have gone all-in on open source products are in the minority while everyone else awaits further refinement and perhaps standardization of underlying technologies. Open source storage solutions may support critical business workloads, but it make time for them to gain mainstream acceptance.

"It is still at a very early stage," Gartner analyst Jie Zhang told TechTarget. "Most end-user enterprises can't yet leverage open source storage software at large scale or for business-critical workloads. But there have been successful cases that serve as both inspiration and proof of concept."

Do enterprises choose open source or use it by happenstance?
Given the amount of technical expertise required to set up DIY OpenStack-based cloud infrastructure, it's no surprise that many organizations are instead introduced to the platform through a vendor's product. For example, solutions providers such as Mirantis and Red Hat have helped many companies set up enterprise-grade private clouds that run on open source software.

Looking back, this trend of using open source via commercial solutions rather than DIY work resembles enterprise reception of OpenSolaris and its ZFS file system/volume manager. Sun Microsystems made a splash last decade by introducing appliances that utilized industry-standard hardware and a stack with significant amounts of open source software. After that, vendors such as Nexenta, a Seagate Cloud Builder Alliance, took up ZFS and built products that leveraged it to support intelligent, automated utilization of flash storage.

Currently, enterprises have many options for using ZFS, but instead of doing everything from scratch, they can rely on products that integrate the technology. Solutions such as NexentaStor, with open source components, have boosted the performance of cloud storage systems while lowering hardware costs. Companies can avoid adding lots more legacy storage simply to lower I/O latency, while benefiting from the smart use of flash to meet capacity, speed and cost requirements.

While ZFS has come a long way, OpenStack may not have reached the same maturity yet. There have been notable adopters so far, including cloud storage providers, universities, hyperscale operators and institutions such as the San Diego Supercomputer Center, which uses OpenStack Object Storage in tandem with commodity hardware to achieve both flexibility and low costs. However, some of these same proponents have avoided open source storage software due to what they perceive as a lack of refinement.

Buyers looking at the open source solutions are more accepting of ZFS and SAM-FS, at least the commercially supported versions, which makes sense due to the history of these technologies. Lack of support options is one reason behind the lagging uptake of OpenStack and open storage technologies, which so far has been largely limited to expansion of NAS infrastructure.

Progress for OpenStack, Hadoop and open source vendors
Despite the appeal of open source projects, a lot of enterprises have stuck with a single vendor to provide integrated solutions. They may do so to keep things simple, or because they haven't found a commercial open product that meets their requirements.That's not to say that open source hasn't picked up substantial momentum. It has, and it may eventually become a firmer fixture of enterprise IT.

The OpenStack Foundation has noted that since the Austin release in 2010, the project's code base has grown from 30,000 lines with 20 contributors to almost 2 million lines from more than 1,600 contributors. Many startups have created solutions based on Hadoop, and the list of major vendors that offer supported versions of OpenStack Swift continues to grow and now includes HP, Rackspace, SUSE and SwiftStack, a Seagate Cloud Builder Alliance partner.

A 2013 survey from the OpenStack User Committee and Foundation found that there were 387 OpenStack deployments around the world and that storage and backup was the sixth most popular workflow for these implementations. Eight of the deployments stored more than 1 million objects, and 22 were using more than 100TB of block storage.

Interest in Hadoop may also drive open source uptake. Enterprises may be able to integrate more services and intelligence into storage processes, building upon the successes of software-defined technologies.

"The concept of programmable storage is one that I think will evolve and catch on," Evaluator Group senior partner John Webster told TechTarget. "The evolution is toward data stores, which implies data services and intelligence defining what storage is as opposed to just a place you put things [in] and take things out of."

Perhaps most promising is ongoing support of open source vendors by venture capitalists. Writing for Re/code, Index Ventures partner Mike Volpi argued that open source products were ideal for handling increasing complex systems and that it was easier than ever to establish a reliable business model using these solutions. That may be the case, but it will likely take time for enterprises to completely buy into these products for storage and computing.