I never could get Nix working but maybe someone will
*but relies on IPFS.
Useless.
So I dug into the source code a bit to see how it’s used. It turns out that IPFS might actually optional, as per the log line on https://github.com/hyprspace/hyprspace/blob/master/p2p/node.go#L213 (“Getting additional peers from IPFS API”)
The list of required bootstrap peers is hardcoded in the same file, but a few lines above, specifically at https://github.com/hyprspace/hyprspace/blob/master/p2p/node.go#L181
I say might be because - while the required bootstrap peers include a bunch of ones based on bootstrap.libp2p.io - there is a long list of hardcoded ip addresses and I don’t recognize any of them.
So those might be libp2p.io ip addresses, but they might also be IPFS ip addresses, or even belong to someone else altogether. (Edit: There are WHOIS tools online like https://lookup.icann.org/en that can be used to look these up and figure out who they belong to if you are really curious, but I can’t be bothered to do that right now.)
In any case, it looks like the way this works is that from a peer, libp2p tries to look up additional peers, and so on. So at most IPFS would be used as a way to get a listing, but once the desired peer is found, IPFS is cut out of the picture for that particular connection and NAT hole punching is used to establish a direct connection between peers instead (as per the linked wikipedia article, https://en.wikipedia.org/wiki/Hole_punching_(networking )
Do you have more information? Haven’t looked into it for a while. What happened?
Is IPFS something your family and friends check on regularly? I don’t even know what it is.
Considering your reluctance to give any information about your assertion that such a project using it becomes useless, I’m not sure you know what it is either :P
Interesting, it’s on AUR, I will try it.
So it doesn’t need any port forwarding, and works on CGNAT? How the “NAT hole punching” works? Both clients connect to something on IPFS?
Afaik, for DHT with torrent, clients need to know at least one tracker, what is the “tracker” here? Something on IPFS? Who am I sending my IP addresses?
How much overhead does this add to speed? I love with Wireguard, that it’s barely noticeable, really close to p2p speeds, OpenVPN was awful in this regard.
First off great find. I didn’t think to check the AUR. I personally wouldn’t use it as that version is 3 years out of date but its existence means that it might be entirely possible to get a non Nix version. I’m not sure I fully understand why it needs Nix OS but what do I know.
It is all libp2p magic
There have been lots if talks on libp2p and Nat traversal. I suggest you check them out. How it actually works is pretty complex and requires someone more knowledgeable than me to explain. One way it works is that both devices start a TCP connection at the same time which gets the proper ports to open up.
AUR packages ending with"-git" or “-svn” always pull the latest commit from source. The version number means that was the last time the packager had to change something on the PKGBUILD script, not the actual version which would be installed.
Where should I look? Where were these talks? I’m interested.
Edit: I found the whitepaper about hole punching: https://research.protocol.ai/publications/decentralized-hole-punching/
It says it connects to a “Hole Punch Coordination (DCUtR - Direct Connection Upgrade through Relay)”. So for NAT traversal to work, you need a third party, this relay. As I expected. I guess you can self host this, but than you could just host a wireguard server. I guess if you are on a locked down network where you cannot connect to any relay (e.g. how the Chinese Great Firewall works technically they could block it) you can’t initiate a connection behind a NAT.
Nonetheless it seems interesting, but no magic here. Maybe the big difference that the relay servers are distributed, so no central authority to block easily.
That doesn’t match my experience with AUR at all. Usually it pulls a specific git revision and checks the hash. This also ensures that the build shouldn’t suddenly fail to some extent.
Though it’s entirely possible that it’s not like this for all packages, though I find it kind of counterintuitive since your package manager wouldn’t know when to perform an update in this case.
The PKGBUILD looks like it is just building via go. I’m not sure how you would configure it without Nix. I’ll try building it.
What about Tailscale? I know it’s Proprietary software, but still.
Tailscale… is not that good. The underlying wireguard is robust, but tailscale control plane is completely proprietary, as well as their DERP servers that it too often uses completely needlessly. They can also block you off from downloading it, updating, or logging in, if you happen to be in a wrong country.
I’m myself looking for an alternative to it, but having trouble finding something I could share with non tech savvy friends while not being as complex on my end as, say, open/strongswan ais. Any suggestions welcome.
https://netbird.io/ maybe?
Have you considered having Headscale on a cheap VPS? We are actually doing that and it is pretty capable. IIRC, you can configure not to use the tailscale servers at all, and use your own public VPS for coordination. Bonus point, tailscale hired the Headscale developer and maintainer, and they are allowed to work on Headscale while on their payroll. The team looks very much into FOSS.
Yep. That’s the number one contender. Well right after overriding default DERP’s with my own VPS machines. I’ll definitely try it out over some weekend.
One of my other concerns with this and other solutions suggested is the reliance on wireguard which can be subject to fingerprinting and censorship. Do you happen to know if it’d be possible to swap out Headscale’s implementation of wireguard to amnezia? I’ll have to do my homework anyway, but who knows, maybe there are some pitfalls to avoid.
Headscale worked for me, but I get the non-tech saavy friends part doesn’t quite jive with it as a solution.
Still, anyone wanna ditch Tailscale and only use it for hosting sites across proxies? Headscale is great.
YAML?? (╯°□°)╯︵ ┻━┻)
what:
is:
your:
- problem
- with:
YAML
# At least you can have comments unlike in json. Who need comments in a config file anyway.
Nothing too major about how it’s usually used, but the yaml spec does allow arbitrary code execution when parsing a file and relies on the parser to have that feature disabled: https://en.m.wikipedia.org/wiki/YAML#Security
That’s why for python, yaml.save_load()
is a thing. That’s fine for your local config files and may even be a feature for you, but it shouldn’t be used to exchange information between services.
My general view is similar, yaml is better if it should be written by humans, json is better if it should be written and read only by a machine. but hyprspace uses json for configuration, so I don’t really understand cellardoor’s comment