In 2011, Facebook (FB) was besieged by surging data-management costs. It was growing too fast. So it decided to change the way it did data centers.
That change was radical. Jonathan Heiliger, then Facebook’s infrastructure leader, thought it would be a cool idea to open-source what the social networking giant was doing with its own data center hardware. He called it the Open Compute Project (OCP).
This was not a mainstream idea. Most companies, even big technology ones like Facebook, bought their gear from the likes of Hewlett-Packard (HPQ), Dell and Cisco Systems (CSCO). They linked together servers, switches and cables, and then they prayed the combination would be enough to meet their growing demands. It was kind of a crapshoot. Heiliger’s idea set that on its head.
In theory it made sense. Open-source Linux already revolutionized enterprise software development and Android was doing the same for mobile. The idea was to bring the same level of innovation to data center hardware, where products had become costly and oddly, energy inefficient. Plus, because Facebook didn’t compete in hardware, it wasn’t really giving away competitive advantages.
Free Preparedness Kit!
To help you get ready to take full advantage of the bull market of a lifetime, I want to send you a complete Dow 31,000 Preparedness Kit — five distinct free reports! The first free report spells out step-by-step what you must do now to position yourself for amazing profits (and protection) over the next two years. Click here to download now!
Facebook’s initial efforts were met with skepticism. In addition to getting pushback from the legacy hardware makers, FB found leading cloud-computing firms – like Amazon Web Services (AMZN), Microsoft (MSFT) and Alphabet (GOOGL) – were less than supportive. After all, they were trying to discourage Fortune 500 companies from building their own data centers in favor of the public cloud.
Yet, one by one, the pieces began to fall into place for OCP. Early on, it developed an alliance with Intel (INTC), a key player in the development of Linux. Standardizing chip design ensured modularity and made it possible for companies to share. Working with Frank Frankovsky, an ex-Dell hardware designer, and Don Duet of Goldman Sachs (GS), they set up the hardware building blocks and provided OCP an inside track with financial firms. After that, it became a movement.
|A Foxconn factory – thanks to OCP, the Japanese company partners with Hewlett-Packard|
And big companies followed. Apple (AAPL) and Microsoft have committed to OCP. Even Hewlett-Packard is involved, working with Foxconn (FXCOF) to build OCP boxes that are being shipped to data centers all over the globe. Financial-services firms have been the most fervent early adopters. Goldman Sachs has committed to 80% of all new servers being OCP and Fidelity Investments recently revealed how open-source hardware is driving significant cost savings. In a sector that spends almost $200 billion annually on technology and data management, this is a powerful motivator.
OCP is not likely to change the movement to cloud computing. AWS, Microsoft, Alphabet and smaller private-label data centers are building big businesses in the cloud with substantial cost advantages. If anything, OCP helps that effort by standardizing and reducing the costs of hardware. The losers are legacy computer hardware firms like Dell and especially Cisco. The days of bespoke equipment, lavish sales conferences and high margins are over.
Open-source is popular with engineers because it’s the foundation of science. Although it’s not great for profit margins (especially if you’re selling) it’s a boon for innovation. Facebook is the investment winner here.