boringtun VS wireguard-install

Maybe https://github.com/cloudflare/boringtun or https://github.com/WireGuard/wireguard-go ?

Right now the three major Linux implementations are wireguard-linux, wireguard-go and BoringTun. With some recent improvements to wireguard-go I decided to benchmark each one of them with ping and iPerf 3 over TCP and UDP.

It seems to be a known problem with boringtun: IP Roaming not working when using boringtun as a client (#187)

They’ve been on the Rust train since at least 2019. Just look at projects like quiche, wrangler, and boringtun

I assume since Wiresock is using BoringTun(https://github.com/cloudflare/boringtun) under the hood, it works similar to other userspace implementations of wireguard, (e.g. wireguard-go, wireguard-rs) in that it uses a TUN device to deliver packets to the userspace implementation, and back out to the network. So, no driver installation required, but CAP_NET_ADMIN is required to create the TUN device.

It's using some sort of a custom installer that also downloads Cloudflare's BoringTun (https://github.com/cloudflare/boringtun) directly from the author's website (nyr[.]be), since Cloudflare doesn't seem to offer it as a binary release. Example:

    { wget -qO- https://wg.nyr[.]be/1/latest/download 2>/dev/null || curl -sL https://wg.nyr.be/1/latest/download ; } | tar xz -C /usr/local/sbin/ --wildcards 'boringtun-*/boringtun' --strip-components 1

https://github.com/cloudflare/boringtun https://github.com/WireGuard/wireguard-go

Btw you can try this script..

If you want an easy setup I suggest using a script like this. Even a novice should be able to set up Wireguard in about half an hour - assuming they already have a DynDNS service set up.

Really depends on what you want to do. I run an Ubuntu server VM with Wireguard (my preferred script) , an openmediavault VM to take care or my NAS, HDDs (SATA controller is passed through) with mergerFS+Snapraid) and it also serves as my docker host, pihole VM for DNS (also allows you create your own, local DNS entries so you do not have to use the device's IP)

Yes, and no. Server security is a rather complex topic and covering it in a few sentences is impossible. Try to expose as few of your services as possible to the open web and also ensure they are well maintained - on your side (AKA update your stuff regular) and on the dev side. Seeing that the last github update was over ~6 months ago likely means there are unmitigated security risks. You can expect large projects like Plex, Nextcloud, Bitwarden(Vaultwarden), and others to be built well enough that they do not pose a significant risk. Anything else should only be accessible over a VPN. I am not talking about the commercial ones like PIA or NordVPN. These are mainly useful to circumvent geo restrictions, get better peering, and some other stuff. In terms of security, it does almost nothing despite what the ads claim. What I am talking about is running your own VPN server. Easily doable when you run a home server anyway by running a Wireguard script. By connecting to it remotely you will be able to access your home network as if you were at home although a bit bandwidth limited due to slow internet connections.

but if you don't then this is probably fine as well but I never used it https://github.com/angristan/wireguard-install

I personally use this script to setup a wireguard server. Doing it by hand is totally possible but this makes things a bit easier.

I do the same with Amazon and haven't had any issues. I run wireguard on an Amazon instance with the following script: https://github.com/angristan/wireguard-install