Google has shared more details of its efforts to implement BeyondCorp a security approach the company has been using internally for years to control access to enterprise systems and data.
In a blog July 2, Max Saltonstall, technical director of Google’s office of the CTO said the company was releasing the information so other organizations could learn how to implement a similar access control model. Google has previously released multiple research papers describing BeyondCorp and its so-called zero-trust model for user access control.
This week’s blog outlines the measures the company had to take initially to get started on the effort. Saltonstall said Google was releasing the information in response to organizations that had reviewed the company’s earlier research papers and wanted advice on how to get started on implementing the model.
“They’re looking for step-by-step help in applying these context-based access practices in their particular organizations, so we’ve created a series about some of our best practices at Google,” he said.
Google began work on BeyondCorp in 2010 shortly after China-based threat actors broke into the company’s network and stole intellectual property. Following the attacks, Google began moving away from access control approaches based on the notion of trusted and untrusted networks and practices such as the use of secure Virtual Private Networks (VPN) for remote access to applications.
Under BeyondCorp model, access control decisions are no longer solely based on whether a user is requesting access to an application from inside the corporate network or outside of it. In other words, a user seeking access to an application from inside the corporate network is viewed as just as untrustworthy as a user seeking remote access.
With BeyondCorp access decisions are made based on detailed, up-to-date knowledge about the user, their job roles and the security status of the devices seeking the access. Google has said such a zero-trust model is essential because network security controls alone can no longer be trusted to provide the security needed to protect enterprise applications and services.
According to the company, by shifting access controls from the network perimeter to users and their individual devices, organizations can enable employees to work from anywhere without the need for a traditional VPN or similar secure channel.
“The first step to moving from a privileged corporate network (usually with a VPN at its core) to a zero-trust network is to know your people and know your devices,” Saltonstall said this week.
To do this, Google had to restructure job role hierarchies and redefine job classifications in order to more accurately capture what access levels people in different roles actually required on a daily basis. “We had to answer some tough, but very logical questions like: ‘who needs to see internal bug information; who needs access to source code; who needs to track customer relationships,’ ” Saltonstall said.
In order to implement a zero-trust model, organizations need to also have complete visibility of all the devices on their networks. In Google’s case, the company had to create a new master inventory of all its devices, he noted.
Like many organizations, prior to 2010, Google used a slew of asset tracking and management tools to keep track of its device inventory. For BeyondCorp, Google built a meta-inventory service that pulled in data from the company’s collection of asset management tools and created a central and trustworthy record of all its devices.
Creating the inventory service took time and considerable investment. But it has given Google much better visibility over the devices on its network, what each device might be doing and whether they have security features like required patches, antivirus software and other characteristics.
In addition, in order to deploy a zero-trust model like BeyondCorp enterprises need to understand what applications they use internally and what security policies govern access to these applications. They need to understand job roles, decide who gets access to specific services and put identity-aware security controls in place to govern access, Saltonstall said.