Google Video Offers Glimpse Inside Its Data Centers

Google opened the curtains ever so slightly on its data center operations, offering a look at part of how it protects customer data via some of its physical security, data protection, and storage processes.

Google introduced those processes in a seven-minute YouTube video it uploaded last week, and which it started promoting on Friday.

In the professionally-produced video, Google, via an unnamed narrator, explains that security considerations prevent the company from opening its doors for physical tours, or from showing all the technology it uses to protect customer data.

Instead, the video is Google's attempt to explain three parts of its operations, namely the physical security of its data centers, how it protects customers' data, and how it ensures reliable operations.

Google in the video said it builds its own servers, thousands every year, to its own custom specifications. Those servers, which are used exclusively in its own data centers, come with a stripped-down version of the Linux operating center.

"This helps provide a computing environment which is much less prone to vulnerabilities," the narrator said. "And it also reduces the need for Google Apps customers to deploy their own patches."

Google policy is to not allow public access to its data centers, according to the video. To prevent unauthorized entry, physical barriers are set up at guard stations staffed 24 hours every day by security guards, and the perimeter fence is backed up with large numbers of surveillance cameras as well as thermal imaging devices.

These are backed up by full security teams which respond to potential breaches of the perimeter, and can call on local law enforcement officials as needed, the narrator said.

Access to the facilities is granted through badges which use lenticular printing, which prints 3D images that change according to the angle of the viewer, and is backed up by biometric security technologies such as iris cameras, according to the video.

To help ensure the reliability of customer data, Google stores data in multiple locations, and modifies it so it cannot be read by unauthorized users. "The files that store the data are given random file names, and are not stored in clear text, so they're not humanly readable," the narrator said.

All the hard drives used by Google are "rigorously" tracked throughout the company, and those that either fail or start to exhibit potential problems are brought to a central location for testing.

"If the hard drive does not pass these tests, it's removed from the circulation," the narrator said. "The data on the hard drive is then overwritten to help ensure that no customer data remains on it. The data override is then verified with a complete disk read. This process helps ensure that there's no trace of customer data remaining on the hard drive."

Hard drives no longer needed by Google are then destroyed in a multi-step process to ensure their data can never be accessed, the narrator said. The first step is to use a machine called the "crusher," which includes a steel piston which slowly punches the center of the drive to cause massive deformities of the drive's platters. The drives then go through a shredder which breaks them up into tiny pieces, which are then sent out for recycling.

"No one will be likely to get any of Google's customer data from these drives," he said.

All data in the Google data centers is backed up on tape libraries, the narrator said. "This provides a level of redundancy to help safeguard its customers' data," he said.

NEXT: Keeping The Power On

Google's data centers are also protected with fire suppression technology, as well as the ability to automatically and seamlessly shift customer operations to another data center so customers can continue to work uninterrupted, the narrator said.

The data centers are also backed up with emergency backup generators which power operations until main power is restored.

Main power for the data centers was designed with energy efficiency and sustainability in mind, the narrator said. "In 2009, Google announced that its data centers currently met energy efficiency levels that the U.S. Environmental Protection Agency projected only the most state-of-the-art data centers would possibly meet by the year 2011," he said.

That investment has paid off. Greenpeace International on Thursday released a report on the use of clean energy by major brand companies' IT data centers, and Google received the second-highest rating, bested only by Yahoo!.