VPNs, or virtual private networks, are an important part of any security and privacy toolbox.

VPNs are essentially encrypted connections between two or more devices that enable you to route data through a secure “tunnel.” Companies use them to allow employees to access corporate networks from outside the office. Commercial VPN services try to protect your internet traffic from eavesdroppers by routing it through remote servers. In theory, that means that a hacker eavesdropping on public Wi-Fi or your home broadband provider can’t see what you’re doing online. Routing your traffic through a remote server can also make it look like you’re in another place, allowing people in countries like China and Russia to access sites that are blocked domestically.

But VPN connections are only as secure as the software that underpins them. Security researcher Thomas Ptacek says his industry is generally distrustful of VPN software. “There’s always a gnawing feeling in the back of our skulls” of an unknown security weakness in VPN software, he says. One reason for that is that most VPN software is incredibly complicated. The more complex a piece of software, the harder it is to audit for security issues.

Many older VPN offerings are “way too huge and complex, and it’s basically impossible to overview and verify if they are secure or not,” says Jan Jonsson, CEO of VPN service provider Mullvad, which powers Firefox maker Mozilla’s new VPN service.

That explains some of the excitement around WireGuard, an open source VPN software and protocol that will soon be part of the Linux kernel–the heart of the open source operating system that powers everything from web servers to Android phones to cars.

WireGuard, created by security researcher Jason A. Donenfeld, is smaller and simpler than most other VPN software. The first version of WireGuard contained fewer than 4,000 lines of code–compared with tens of thousands of lines in other VPN software. That doesn’t make WireGuard more secure, but it makes it easier to find and fix problems.

WireGuard clients are already available for Android, iOS, MacOS, Linux, and Windows. Cloudflare’s VPN service Warp is based on the WireGuard protocol, and several commercial VPN providers also enable users to use the WireGuard protocol, including TorGuard, IVPN, and Mullvad.

Building WireGuard directly into the Linux kernel, the core part of an operating system that talks directly with hardware, should make it faster. WireGuard software will be able to encrypt and decrypt data as it’s received or sent by the network card, instead of passing data back and forth between the kernel and software that runs at a higher level.

WireGuard isn’t officially “finished” yet. Donenfeld expects the official release in a few weeks, which should open the door to wider use by VPN providers. Jonsson expects adding WireGuard to the Linux kernel will make it useful for securing connections between Internet of Things devices, many of which run on Linux.

Lessons From Consulting

WireGuard grew out of Donenfeld’s security consulting work, much of which involved what’s known as “penetration testing.” In other words, he got paid to figure out ways to break into companies’ networks. He created the software that eventually became WireGuard as a data exfiltration tool–a way to quietly and securely transfer data off a target’s computer.

He moved to France in 2012 and, like many VPN users, wanted a way to access the internet as though he were connecting from the US. But he didn’t trust existing VPN software. He eventually realized he could use his exfiltration tool to route his traffic through his parents’ computer in the US. “I realized many of the things I’d been doing for offensive security were really useful for defensive security,” he explains.