Skip to main content Skip to Footer

March 19, 2014
Harnessing Hyperscale and the Future of OCP
By: Krista Schnell

Have computer? Can code. It’s this relatively low barrier of entry to the world of software that has enabled “kids” such as Mark Zuckerberg to be so successful in the world of technology. Well, a certain part of the world of technology. Because we haven’t seen this phenomenon in hardware—largely because there’s no equivalent in the hardware space for what a computer did for software (at least not yet—3D printing might be the solution, which you can read more about in the Digital-Physical Blur). Right now, between the cost of materials and manufacturing equipment, it’s just too hard to get started.

So it makes sense that software has had its spot in the limelight for a while now. But that doesn’t mean there hasn’t been progress in the world of hardware. In fact, as software continues to improve, it places new demands on hardware. Not only that, but because software and hardware are so dependent on one another, improvements in software create opportunities for improvement in hardware that have never been there before. This means the likes of Amazon, Google, and Facebook—who are perhaps thought of as software giants—are hardware giants, too. They have been working on hyperscale systems (supersized, super-scalable, and resilient data centers) for years now in order to support their “always on” and “always optimized” businesses. But the hardware innovation they’ve fueled presents opportunities for everyone, even those who aren’t “hyperscale” (yet). Accenture’s 2014 Technology Vision highlights this trend and more in the chapter Harnessing Hyperscale.

One of the best examples of hardware innovation being made available to others is through the Open Compute Project (OCP). Bringing it back to Mark Zuckerberg, he realized that Facebook needed hyperscale systems but could not compete with Amazon and Google, who had already created their own proprietary systems. Facebook felt that if they could leverage the intelligence behind the community, they could rapidly accelerate their infrastructure improvements and even perhaps “blow past what anyone else [had] done,” according to Zuckerberg. And in fact, Facebook has built its data center at one-fifth the cost of a traditional data center; they have saved $1.2 billion dollars and an equivalent energy usage of 40,000 homes. The ultimate goal of OCP is to “build one of the most efficient computing infrastructures at the lowest possible cost,” and because it’s open, it is something that everyone will be able to access and leverage according to their needs.

All of this sounded great until January 27, 2014, when Microsoft announced that they were joining the OCP alliance. Then, it sounded awesome. Microsoft, with their typical business model of proprietary software and paid services, joined the data center hardware market, and it did so in an open way. By contributing designs of cloud servers that run services like Bing, Office 365, and Windows, Microsoft became the first global cloud provider to publicly release server specifications through OCP. They also provided the software they created for the management of hardware operations. So why did Microsoft do this? Well, they wanted to gain ground against rivals like Amazon and Google. They wanted to show their support and involvement in the open source community and accelerate the growth of hardware innovation for cloud computing (which could benefit them, if customers design to Microsoft’s specs and therefore choose to invest in complementary Microsoft products). According to Wikibon co-founder and CTO David Floyer, the effect of this is that Microsoft’s participation “catapults OCP from a contender to a favorite to become the de facto standard for the open source hardware market, particularly for cloud providers and data center providers.”

By Microsoft joining OCP, hyperscale systems and hardware advancements are even more accessible to the larger community. It means that companies who saw Harnessing Hyperscale as a trend in our 2014 Technology Vision and thought it wasn’t for them now need to pay more attention than ever. Hardware innovation is becoming accessible in ways that it never has before. It doesn’t mean that there will be a new Mark Zuckerberg in the hardware space tomorrow, but we’re headed in a promising direction, for large businesses, small businesses, and perhaps even “kids” alike.

Popular Tags

    More blogs on this topic