No comprehensive account of Norse mythology was ever written down by those who believed in it, or if it was, those records or references to them have not survived to this day. The only accounts we do have are those compiled by Christian monks who were converting the old Norse believers, right at the end of the pagan era.
These accounts are heavily colored by the expectations and the worldview of those monks, and we do not know where the ideas that the Norse actually believed in end and where the ideas of what the monks thought religion is supposed to be begin. For example, much of the popular conception of Ragnarok is heavily influenced by Christian eschatology, instead of the original Norse beliefs. To complicate it even further, the Norse beliefs were in no way static, and as the influence of Christianity spread, the beliefs might have morphed to absorb concepts from Christianity.
"For example, much of the popular conception of Ragnarok is heavily influenced by Christian eschatology, instead of the original Norse beliefs. "
In more easy words: christians believe that one day the world really ends with judgment day/apocalypse, and interpreted Ragnarok similar as the end of the world, which is what many people today think of Ragnarok (and that view that gets reinforced by the popular movies)
But Ragnarok is not the end of the world, just the end of a cycle and start of a new beginning. The snake that bites its own tail. Endless cycle of seasons. Really a different philosophy.
> christians believe that one day the world really ends with judgment day/apocalypse, and interpreted Ragnarok similar as the end of the world, which is what many people today think of Ragnarok
There is also the interpretation that we live in a post-Ragnarok world, which conveniently allows the Christian narrative to perfectly mate to the end of the Norse gods’ reign.
Jackson Crawford [1] does a decent job of explaining what bits of the mythos we still have. He also explains some of the misconceptions you speak of and even some of the things that movies get right and wrong. Related to some other discussions in this thread, he also pronounces many words for people.
Maybe but the eddas are 7 and 800 years old respectively, so still older than 600 (and these words almost certainly existed before being written down.)
It is. The written sources we have for norse mythology are 13th century, and those are compilations of earlier (lost) writings and oral traditions. There are written fragments from the 10th century referring to the norse gods.
There are even 6th century writings referencing a germanic mythology with many similarities to the norse mythology.
First of all, it's a branch of Indo-European mythology, as are the Greek, Roman, and Hindu pantheons. So it has existed in some form for thousands of years. But "Norse mythology" as we know it was mostly written down in the 13th century - so 700+ years ago.
Can anyone help with the pronunciation? There's no help from the Wikipedia article. (Not that Wikipedia pronunciation guides are of any use to anyone other than language nerds, anyway.)
In parts of Europe it's quite common (at least in Sweden and Norway I know it's true) to learn the IPA so you can make sense of the alphabet and its phonemes, no need to be a language nerd to have a passing knowledge of IPA.
I recommend you studying it a little, it's really not that hard, just looks weird.
Can confirm. Also learned it in school in Germany, although I forgot most of it, because I almost never use it. For most things, forvo [0] works good enough.
I like the IPA Chart website [0] which lists the phonemes and has examples of each when clicked. To understand a pronunciation (such as /ɪg.drə.sɪl/ for Yggdrasil), I'll match up the symbols to the chart and piece it together from the sample sounds.
Sidenote: I had a bit of trouble locating an IPA transcription for Yggdrasil. The pronunciation guide in the Wikipedia article for Yggdrasil Linux/GNU/X is not IPA, it's using English pronunciation rules to try to emulate the correct pronunciation. The pronunciation in dictionaries (at least Merriam Webster) is also not quite IPA, though it is close. I came up with this pronunciation by merging those sources. I is probably correct, as I found a matching transcription in an old version of the Yggdrasil Wikipedia article which was removed for being original research.
Their Linux CDs proved essential in a time where bandwidth was scarce and expensive with home users on dialup and only universities having a few megabits of bandwidth.
Not only that - the distribution my dad brought home included a _book_ (remember those?) which was actually a bunch of open source books collected in one! Tons of information on Unix stuff, shells, commands etc. I learnt so much from those books. I have no idea how someone would start in this day and age... (this was so important to tell for me that I finally stopped lurking and made an account _just_ for this comment!)
> the root is the node with the lowest ed25519 public key, rather than the highest sha512sum hash of the public key
With this scheme, could a bad actor decide to choose a poor key just to be the neighbor of a target in (edit) keyspace? Ordering by the hash of a public key means that the order is protected by the hash function's preimage resistance; does the generation of a ed25519 key have a similar protection?
My point is that I don't think there's anything during key generation that requires the resulting key to be secure / chosen 'well', so an attacker might exploit key generation as a way to target a particular spot in the line, where having an insecure/easily compromised key doesn't matter to them even though that may be detrimental to the health of the network. But at least with preimage resistance any public key is just as likely as any other to get a particular spot that they desire so there's no incentive to exploit key generation to get it, they might as well generate a bunch of secure keys if they want to use a bitcoin-style brute force strategy.
Can someone explain, why it has end-to-end encryption by default on this level, and why it is good? Isn't this project more about host discovery and routing. Is it providing more performance compared to encryption on other layers, or just for "easy automatic" data encryption?
Based on documentation, it sounds that they have some kind of own crypto implementation in the end.
I found the whitepaper describing used algorithms, but I would need to know more how exactly they are applied and why they are selected, before I could trust the encryption.
This allows us to focus our investment on one layer and have it apply to all applications on top "for free". I'm not saying this specific implementation is the right one, but having it below the app layer makes a ton of sense.
Yggdrasil builds a multi-hop IPv6 overlay network using peer-to-peer connections. You can string a whole bunch of nodes together using whatever means (cables, wireless or TCP peerings over the internet or any other network) and you get a full-mesh network where everyone can reach everyone else.
It's designed to be as minimal-configuration as possible and scalable in a way that many other mesh routing protocols aren't.
Quite different.
ZT is centralized config, direct peer to peer, automatic key exchange, virtual switch.
YG is decentralized, direct peer to peer and multi hop routing, manual key exchange for direct peering, virtual internet (each node is a router to another nodes)
ZT (and Tailscale) requires a central node coordinator for automatic config and peer key exchanges, peers directly to each node to create a flat virtual network, not multi hop routing, between two peers you need a reachable IP or port mapping (supports UPNP) by one of them (fixed in config there are private TCP relays when carrier NAT/double NAT/ restrictive firewalls are in place, very slow), Uses UDP for the links, mimics a virtual switch and support custom IPV4 / IPV6, good for virtual private networks.
YG (and CJDNS) is kinda an overlay public network, is fully decentralized, it supports automatic routing between nodes to mimic a virtual Internet, each node is a router too, needs to register each key for every direct peer node connection (config needs peer key, reachable ip and port, but supports automatic key exchange for local networks), YG uses TCP for the links, support self-healing routing, every peer has an IPV6 address derived by its public key.
ZT (and Tailscale) can mimic a mesh network and node routing because supports bridge nodes between networks (routing between virtual switches), but is not self-healing and somewhat heavy work to config. ZT is fully open source, can be config with your own node coordinator and discovery helpers (Controllers and Moons in ZT), not easy. Only Taislcale client is open source, can't be config with your own node coordinators for free.
You can mix ZT and YG, weird side effects warranted but works.
(Edited post - some grammar fixes)
It's an overlay network (a software defined network that runs on top of another network... like you internet service providers network).
It's closer to bittorent than a VPN. It has end-to-end encryption and each node (the app that runs on your PC) distributes routes to each other (similar to how routing works on the Internet between large networks). It appears to be a flat spanning-tree style network.
They mention that it is using similar code and ideas as the cjdns project.
Do I understand correctly that it does not do hole-punching, and e.g. unlike with a VPN a host behind NAT will not be able to accept incoming connections?
This is correct in the sense of peering, however hosts behind a NAT can simply connect to any other host on the network such as a pubic peer and then they can accept incoming connections over the yggdrasil network.
I use yggdrasil for NAT hole punching my VPN, for example.
I haven't used tailscale so... I think so? Tailscale is pretty light on the details, and even their docs say that they have been reluctant to describe how it works.
>People often ask us for an overview of how Tailscale works. We’ve been putting off answering that, because we kept changing it!
Yggdrasil is fully peer to peer and doesn't require a central coordinator like tailscale does. Ygg is closer to a global network than a private one. You can make a private network, but if any peers on the network peer to the global net then your whole network is now peered. this should be handled at the firewall level, or with an overlay VPN.
it's fairly circular actually - we used Yggdrasil in some of the earlier P2P Matrix POCs. This showed up some limitations, so we worked on solutions for them (alongside Yggdrasil) which became Pinecone on the Matrix side, and Ironwood on the Yggdrasil side. Then Yggdrasil 0.4 is incorporating them for general purpose IP overlay routing, while Pinecone is focusing on Matrix-specific application layer overlay routing. Perhaps the two will converge in the end, but until then they're somewhat overlapping friendly sister projects :)
One thing that wasn't clear to me from the documentation: What's the typical latency you observe with this network? Does the routing take physical distance/observed delays into account in some way, or could you wind up with short (in network space) paths that in reality bounce a packet back and forth between the US and New Zealand repeatedly?
Latency is dependent on the underlying peering connections. Yggdrasil will try wherever possible to take the shortest paths in network space, but yes, it's possible those could be physically indirect if the underlying peerings are indirect. Generally on the public network (which is probably a couple thousand nodes in number, where people have contributed a number of public nodes and have interconnected them) we see very reasonable latencies.
How does Yggdrasil compare to Wireguard? A github search shows that yggrasil-go uses wireguard-tun project as the tun driver; does it relate in any other way? The main problem/use case is different of course (Wireguard is a manually configured point-to-point vpn with e2ee, where yggdrasil is an internet-scale overlay network with e2ee.), but I mean at a low level, protocol, encryption, etc.
At low level is not related to WireGuard in any way.
The wintun project only exposes a virtual network interface to Windows systems, a clean and proper signed windows driver. That side project to Wireguard was created because the native windows virtual interface is too basic for "advanced vpns" and the former windows TUN driver (OpenVPN side project TUN driver for windows) used by everyone was clunky and old.
YG uses it own crypto and routing, wintun is used here only to expose the virtual network interface on Windows.
I see, so there's no real integration between the projects besides the incidental TUN server. It seems like YG would benefit from using WG as a primitive for point-to-point encrypted comms between nodes, and build the spanning tree system on top of it.
I’d also be curious about this. Say for the use case of running a small private chat server hosted on a home network; does either of these seem better suited, or are they just different architectures that can handle mostly similar things?
Very different.
WireGuard is explicit basic, only cares how to make an encrypted link between two devices, and do it very well. Key exchange, IP assignation and routing are manual work. There are solution build on top of WireGuard (Tailscale is one of them) that puts some "automagic".
YG puts more "magic" on protocol (autorouting, mesh making, etc), but is not that clean on design (crypto not formally tested, latency prone TCP links, not good enough NAT punching, etc).
Wireguard and YG are different tools on the SDN network toolbox, and can be mixed for special porpoises.
I have two devices split by VNET and not routed out to the internet. I connect those two and a VPS to create a small Yggdrasil network. This allows me to access all three devices from “anywhere”. Would use again.
Good question. I could achieve the same with Tailscale or Innernet, both using Wireguard. So perhaps my answer has to be that Yggdrasil tickles my interest in mesh networking.
It works pretty similar to cjdns, but implementation is not as crashy in my experience, also some transport protocol differences (ygg uses tcp while cjdns does udp), also DHT routing instead of cjdns supernode stuff.
So while usability is pretty similar, they're pretty different underneath.
I haven't checked, but this may be due to the restriction in CCM/GCM.
Here is the relevant comment/code from Samba.
* CCM and GCM algorithms must never have their
* nonce wrap, or the security of the whole
* communication and the keys is destroyed.
* We must drop the connection once we have
* transfered too much data.
*
* NOTE: We assume nonces greater than 8 bytes.
*/
Nope, is more like a "Vitual public network" maker.
Peer links are encrypted, each peer has its own key, but DHT and routing is not obfuscated. It exposes virtual network interfaces with an IPV6 address per node. You can use common software on it.
"Magic VPN" or "Magic E2EE LAN" kinda IPSEC for commoners, depends on how you config it.
the name reminds me of a paper came across my desk couple of months back: https://arxiv.org/abs/2007.11403 - "Yggdrasil: Privacy-aware Dual Deduplication in Multi Client Settings"
Could someone summarize what exactly the implications are of the exception? It seems that it is a link-time firewall for any GPL viral behaviour? As long as you're on the other side of the linker you can do anything, including embed a statically linked version in a proprietary application?
My guess would be that this allows developers to distribute applications on Apple's App Store without much of a headache, where otherwise the user has no (LGPL-guaranteed) way to swap out libraries otherwise. IANAL, and haven't spent any significant research on this fwiw.
Wow hey there, let me counter that opinion with my opposing that is very fitting. Sure the word might look a bit crazy to non-nordic people (it is kinda crazy, hard to disagree there) but the meaning of the name is imo really well chosen.
It comes from the fairly well-known norse myth of the world-tree spanning all realms. So a overlay P2P network based on a global spanning tree feels like a very fitting name. But hey, you are free to dislike it for personal reasons if you want :)
