BoringTun, a Userspace WireGuard Implementation in Rust

Let's see how it continues, so far it looks like compatibility was an important goal of the Cloudflare team.

Yes, that's unfortunate and I think he's handled it well there. It was the same with the guy who did a windows implementation, didn't want to work with the upstream so is rife for fragmenting the protocol long-term

It's interesting that this seems to mirror our experience with Cloudflare's Rust implementation of QUIC:

https://news.ycombinator.com/item?id=19019150

One of CloudFlare's goals is to have a BSD-Licensed result. Seems reasonable to maintain isolation from the core WireGuard team that is publishing under GPL to support that goal, just to ensure they don't accidentally end up with GPL code in their codebase.

The Go implementation is fine. I don't think you have much to worry about.

But we're pretty frequently running into limitations and fighting with the runtime; we've submitted quite a few random patches to it. I think Rust is probably a better-suited language for a high performance and relatively secure WireGuard implementation. Hence, we'd really like to have a strong first-party Rust implementation for the project.

I have been getting the mio question a lot.

Mio, in theory, is not very far off of what I ended up implementing, but it is missing a few features that I really needed, and building those features on top of Mio would be harder than just using epoll/kqueue directly. Also I don't understand the fear people have of using epoll/kqueue. Those are very basic features that every engineer has to know how to use.

So what are the features:

No locks - Mio polls under lock, but epoll/kqueue are multithread safe by nature as long as you use the correct filters. Mio has considerable overhead over pure epoll/kqueue.

No need for tokens, instead I store closure pointers. Real zero overhead. Of course I could cast pointers to usize and use as Mio tokens, but I would still have to build the ownership model on top, which is most of the work anyway.

Timers in Mio run in a thread??? I use platform idiomatic conventions instead, which with kqueue require direct access to kqueue.

Signal handlers, can't make idiomatic one with Mio (or maybe I missed how).

Same with notifications. Could use something awkward like a channel, but prefer idiomatic approach.

It is still quite expensive, the kernel module is still faster. Performance depends a lot on your kernel version and network card. We measured anywhere from 20% to 40% advantage to kernel module. But we are working on additional performance features, such as using send/recvmmsg and raw sockets.

Currently the library part supports Windows, and it can be used from C# or C++, if you want to use something like https://docs.microsoft.com/en-us/uwp/api/Windows.Networking....

However there is no native support for tunnel interfaces in Windows, and we are still thinking about the best way to add full support.

The WireGuard project for example developed a whole new driver just for that. Don't know if we are going to follow that route.

I am a huge fan of Wireguard and use on all the major production servers I currently manage.

But user management is exactly as sophisticated as SSH key management. That is to say, easy for 2-3 people, a pain otherwise.

There are attempts at creating UI based user managers for wireguard, and I hope those go somewhere.

I bet the major wireguard-supporting VPN providers have solved this problem for themselves though.

Update: Just wanted to add that I don't think that user management (the work of actually provisioning and managing users) should fall under the umbrella of Wireguard. It's perfect the way it is. But I suspect there will be an ecosystem of tools that come up around it.

WireGuard is currently in a place where it works really well for site-to-site VPNs (where network/devops engineers are the only ones interacting with it directly, and it's just magic for everyone else) and for access (laptop-to-VPC) VPNs for engineers.

It's not yet in a place where I'd recommend it for access VPNs for, say, support staff.

It's more pleasant to use and manage than SSH or SSH tunnels (this is surprising and speaks to what an achievement WireGuard is and why people are so nuts about it), so if you calibrate your tooling/slickness expectations accordingly, you should do fine with it.

IIRC, Wireguard doesn't target "undetectability" as a feature.

Wireguard is better than other VPN software because it's intentionally minimal, with a small set of crypto in-use and simple design choices, leading to small easily-auditable codebases. It's also got a pretty straightforward set of config options without a barrage of tunables, which makes it pretty much impossible to subtly misconfigure things in a way that harms security.

You'll always be able to tell VPN traffic from the rest due to behaviour patterns. If you can tell apart Netflix movies from packet sizes/timing, you can easily tell when something's a persistent tunnel for various connections (i.e. VPN).

The only way would be to keep the pipe continuously saturated with some padding while you're not sending real data... But is that worth it?

Wireguard focuses on being a better vpn where better is vaguely defined as using modern crypto in a way that is hard to get wrong. As a result, it is a lot simpler with far fewer bells and whistles. This means it tends to be orders of magnitude faster than an equivalent usage of say openvpn (I measured 15mbps vs 52mbps awhile back with openvpn and wireguard respectively).

What you want is a slightly different albeit very important problem to solve reliably.

Yes. Besides being fast and small, it connects very quickly, does seamless roaming and works without keepalives, making it incredibly well suited for mobile clients.

Easier to set up, faster, safer and it does not restart connection when you go from Wifi to mobile.

Open Issue https://github.com/cloudflare/boringtun/issues/38

According to the blog post:

"The name might sound a bit boring but there's a reason for it: BoringTun creates a tunnel by 'boring' it. And it’s a nod to Google’s BoringSSL which strips the complexity out of OpenSSL. We think WireGuard has the opportunity to do the same for legacy VPN protocols like OpenVPN. And we hope BoringTun can be a valuable tool as part of that ecosystem."

so coincidental.