13 Back in the 90s, I was lucky enough to start my professional life right at the rise of wireless communications. As I was traveling around the globe to design GSM networks, I felt like I was part of a major revolution in the way people could communicate. It was exhilarating and fulfilling. We thought that everything was possible. And it was, as the dreams of the visionaries have since been delivered on our smartphones. But the way we communicate isn’t the only thing that industry disrupted. Wireless networks disrupted the current understanding of telecommunications networking. The major force then was the emergence of softswitches, which enabled more scalable and more powerful telecom network switches with software running on industry standard servers and Linux. For the likes of Ericsson or Alcatel-Lucent, engineering approaches changed radically. No more proprietary hardware. No more proprietary OS. Welcome to the freedom of choice. Welcome to the world of software running services on commodity hardware. Well, we are now in 2014. And the storage industry is the cornerstone of our digital life, with an overwhelming mission of coping with our ever increasing needs to store data, in a more cost-effective and environmentally-friendly way. Well, most of storage systems still rely on appliances, with proprietary hardware and OS. There were good reasons to design storage systems that way 20 years ago. In order to extract performance from disks on an aggregated basis, you had to optimize every layer of the associated software, and control from the storage software all the way down to some very low level functions, such as turning the fan on or off in the disks enclosure. But none of these reasons makes sense today. Because industry standard servers coupled with Linux and software such as the Scality RING simply do a better job at scaling for the Petabyte era, while delivering much lower price points. It was long overdue. But now the storage industry can embrace software on commodity hardware. And this isn’t a vision. It’s reality. By formalizing our partnership with HP, the leader of the industry standard server world, we at Scality demonstrate through our joint 50 petabyte installed base that Service Providers and large Enterprises can enjoy today this state-of-the-art way of building storage systems. We bring to the Enterprise the secret sauce of the Internet giants, who have shown a very clear path forward over the last decade. So beyond the cost reduction, the freedom of choice between hardware and software and the feeling of doing things “the modern way,” what does it mean? As an operational guy I am sensitive to the operational implications. When I got introduced to the Scality RING, one of my first “deep dives” was to look at our support tickets flow. Having delivered storage appliance systems in the past, I knew too well the bad habits of the storage industry: churning hardware to restore service… without really understanding if the root cause of the issue is software or hardware-related. Well, at Scality less than 2.5% of our support tickets involve us in discussing if root cause of a given customer problem is hardware or software related. How could that be that the demarcation between hardware and software issues is so clean? Well because the RING is software that runs entirely in user plane, which doesn’t require any Linux kernel modifications. It takes two hours to train a support engineer to triage hardware and software issues with the RING. Five years of production experience and hundreds of millions of users depending on our product prove it. Another great benefit of our approach is that you can manage the hardware infrastructure the way you like. Want to upgrade a system with new hardware and mix match generations? Easy. Because we don’t even need a certification program. We meet new kinds of x86 Linux servers on the ground all the time. And it works. Because we are basically seen as an application on these servers. Last but not least, we allow our customer to relax their conventional “4-hour Turn-Around-Time” SLA terms on hardware maintenance. Not that disks don’t fail under the RING… they do, as all storage disks. But our software abstracts the hardware failures and was designed with self-healing in mind from inception. So why wouldn’t you show up only once a week in the data center to replace failed disks vs. worrying about it 24×7? This announcement with HP is a critical step. The leader of industry servers partners with the leader of true storage software. And this is not slideware but the formalization of a relationship which has pleased many customers already, building off a 50PB installed base. The time has come. Welcome to freedom of choice in the storage world.
