Facebook Rapidly Expanding Data Center Footprint And Technology


The social media giant is doubling down on data center investments in 2018 through new facilities and technology to meet increasing data demands.

"The most successful companies in the future are going to be the ones who have the data and how machine learning can distribute that data to discover new insights," said Zeus Kerravala, principal analyst at ZK Research. "Facebook is, frankly, prepping for the future."

At the 2018 Open Compute Project (OCP) Summit this week, Facebook revealed that it would be expanding the amount of data centers in Papillion, Nebraska from two to six buildings.

[Related: Michael Dell: On-Premises Solutions Are More 'Cost Effective' Than Public Cloud For 85 To 90 Percent Of Workloads]

Sponsored post

Menlo Park, Calif.-based Facebook also confirmed last week that it would be building its ninth U.S.-based data center in Newton County, Georgia, expected to be a 970,000 square feet facility opening in 2020. The social media giant is investing $750 million initially in the project, planning to spend a total of $2.5 billion on the facility over the next decade.

Since 2017, Facebook has announced five new cloud campuses with a total of 12 data centers across the globe. Facebook opened the doors of its first data center facility in Prineville, Oregon in 2011.

"Facebook is putting more data closer to where the users are, and that makes a big difference when you're doing streaming video and things like that," said Kerravala. "We're seeing a shift in the market away from these big data centers more toward distributed data centers. With so much artificial intelligence and machine learning being done, you want to move the data back closer to the user, so you want to build distributed ones."

On Tuesday, Facebook unveiled its new Fabric Aggregator solution, a new system to manage data traffic between the company's data centers, according to a release. At the core of the new solution is Facebook's Wedge 100 switch. The new networking system was needed to support its growing data needs, according to a blog post by Facebook's Engineering Team on Tuesday.

"Using a large, general purpose network chassis no longer met our needs in terms of scale, power efficiency, and flexibility," said the blog. "Taking a disaggregated approach allows us to accommodate larger regions and varied traffic patterns while providing the flexibility to adapt to future growth."

Facebook executes said it would donate its Fabric Aggregator design to the OCP, which the company co-founded in 2011.