The OCP Summit V, an event that took place at the end of January at the San Jose Convention Center in California, gave the audience the opportunity to witness a “Fireside Chat” between Tim O’Reilly, Founder and CEO of O’Reilly Media, and Mark Zuckerberg, Founder and CEO of Facebook.
For some of the discourse, Mark has pretty much reiterated what Jay Parikh, the company’s VP of Infrastructure, has stated during the keynote regarding Facebook’s history, breakthroughs and vision for the future.
“In terms of why we’re interested in OCP, Facebook has always been a systems company,” explained Zuckerberg, “and a lot of the early work was around open source and building systems. As a matter of fact, a lot of the early social networks had trouble scaling.”
Living up to the challenge
“As a company, we took those challenges really seriously and that’s part of the reason why early on we were able to develop and deliver such a fast and efficient service,” said Zuckerberg, who added that, as the company grew and got to a bigger scale, it prompted them to look for ways of doing things more efficiently. “That’s when we started thinking about designing our own hardware, software and systems.”
As O’Reilly noted, Facebook was not the first company to design their own hardware (Google was famous for doing that years ago), but they were the first to open it up to the community. “What motivated you to do that?” he asked.
“A number of factors,” admitted Zuckerberg. “From an idealism perspective, we want to bring the technology and the innovation that we are designing to the world.”
Zuckerberg highlighted the key factors of Jay Parikh’s presentations during the event: greater efficiency and money saving strategies. He explained: “In the last three years alone Facebook has saved over a billion dollars in building out our infrastructure using open compute designs.”
Some really intimidating figures
“We’ve also saved a lot in terms of energy consumption and emissions,” bragged Zuckerberg. “Just in 2013 alone we’ve saved the equivalent of energy required by 40,000 homes per year and the amount of emissions corresponding to taking 50,000 cars off the road for a year.”
Going back to their interest in open source and open compute, Zuckerberg elaborated: “When you’re the first company to design something, sometimes there’s an advantage in keeping it proprietary and secret, but when there are other companies who have done the same sort of work it’s better – in our opinion – to collaborate and work together with the community.”
“That’s been the story for open source all along,” noted O’Reilly. “It’s an incredible efficient way of doing an industry leapfrog.”
“However,” he continued , “a lot of people thought it wouldn’t work for hardware. What made you think you could pull it off?”
Giving it a bit of context, Zuckerberg explained: “People are spending so much money in building up infrastructure; therefore, saving money and improving efficiency is creating a strong incentive. We’ve tried to form partnerships instead of doing everything ourselves.”
As far as open compute was concerned, “it seemed like we should try to do the same on the hardware side.”
New bold projects
“Now, with the success of Open Compute, we’re using it as a model for another big initiative of ours: internet.org, which is trying to bring affordable – or free – basic internet services to everyone in the world.”
After trying and achieving the milestone of connecting one billion people, the next big thing for Facebook is building a network that will help connect everyone.
“What exactly are you doing to accelerate that?” inquired O’Reilly.
Zuckerberg admitted in being challenged to bring the costs down: “People keep stating that there are over 5 billion smartphones out there. But that’s not all. The most expensive part of owning a smartphone is its connectivity. The network and the access data costs three times as much as the phone itself.”
Therefore, Facebook’s approach is using an Open Compute-style partnership in the industry coalition in order to try to make the network components of the infrastructure cheaper (that means licensing, spectrum, all the different things that go into delivering the internet).
“We try to use our position and work with carriers around the world to craft a business model where they can offer basic services for free (email, text, weather access, search, wikipedia),” added Zuckerberg.
“A strategic thread in the case of Open Compute,” noted O’Reilly, “is that we’re going to use more and more server capacity over time, so we’d better make it cheap and efficient.”
As Zuckerberg said, only one-third of the world population has access to internet at the moment.
He believes that “bringing the cost of the overall network down and delivering the internet for more people will have a very profound impact on the world.”
The shareholder value myth vs. idealism
O’Reilly felt the need to ask Zuckerberg how much of the shareholder value myth was actually true: “it is believed that the only obligation of a company is to make money for its shareholders,” he said. “But some of the things you mentioned here are driven by idealism. What do you think is the responsibility of a company?”
Zuckerberg repeated his strong interest in the energy efficiency and in the future of the knowledge economy.
Asked if every CEO should be thinking like that, he replied: “There aren’t many companies in the world who have the resources and the reach that Facebook has. That means we have more responsibility to do more of that kind of work.”
The right thing to do
“There’s no real business plan that’s going to support the billions of dollars we are about to invest in the internet.org trying to bring connectivity to the rest of the world, but we’re doing it because we think it’s the right thing to do and I think it has a way of catching up in a positive way over time,” added Zuckerberg.
“A lot of the things people do are have done are because they have a passion for it before there’s a clear business case. But if you’re doing it good, there is always a way to make profit out of it, later on,” believes Zuckerberg. “It’s one of the tricky ways of the capitalism: you can’t connect all the dots ahead of time.”