Yggdrasil Network

Original Article ↗ Hacker News Discussion ↗

Project maturity and activity

  • Some initially assumed Yggdrasil was abandoned due to sparse blog updates and having first seen it years ago.
  • Others, including a core developer, state it is active but a free-time project, with v0.5 (new protocol design) and 0.5.9 (latency-focused link cost changes) released recently.
  • A blog update is requested to better reflect development progress; some argue code activity is a better signal than blog posts.

Real‑world use and benefits

  • Several use it as a mesh VPN to reach home servers or personal devices behind NAT, including over IPv6-only or locked-down consumer ISPs.
  • Multicast peering on LANs is praised: devices can talk directly using the same Yggdrasil IP whether local or remote.
  • Compared to configuring WireGuard, dynamic IPs, and port forwarding, setup is reported as simpler for some scenarios.
  • Performance via QUIC peers is reported at ~80% of WireGuard in at least one anecdote.

Technical design and limitations

  • It is an overlay network: virtual IPv6 mesh on top of existing IP, with public-key–derived IPv6 addresses.
  • Address truncation to fit IPv6 raises collision concerns; currently considered acceptable for a test network, with plans to move to a custom protocol without length limits.
  • Earlier designs used a DHT; v0.5 removed it. A partial spec exists but is outdated; documentation is intentionally lagging while the design is still evolving.
  • Links currently assume reliable, ordered transport (often TCP), which complicates NAT hole punching and encourages routing traffic through peers rather than direct multi-hop links.

Comparison to other systems

  • Positioned closer to an experimental replacement for BGP/mesh routing than to a turnkey VPN like Tailscale or WireGuard.
  • Alternatives mentioned: Tailscale/Headscale, Nebula, Reticulum, Meshtastic, cjdns, NKN, Irdest, telehash, and others; these cover overlapping but not identical goals (VPN, key-addressable overlay, mesh over non-IP links, etc.).

Naming and usability

  • The name “Yggdrasil” provokes debate: some see it as hard to spell/pronounce and a barrier to adoption; others value its distinctiveness and cultural reference.
  • There is mention that the protocol might eventually be standardized under a different, more neutral name.

Anonymity and scope

  • Yggdrasil explicitly does not aim for anonymity, likened more to BGP/OSPF than Tor.
  • Some view lack of built-in anonymity as a deal-breaker; others argue anonymity should be layered on top to avoid Tor-like performance penalties and allow low-latency routing.

Broader networking and mesh ambitions

  • Discussion broadens to whether a worldwide P2P mesh could replace the IP layer.
  • Consensus in the thread: technical protocols are not the main barrier; physical infrastructure, cost, and scaling issues make fully decentralized, ISP-free global mesh networks impractical today.
  • Yggdrasil can in principle run over non-Internet physical links, but participants emphasize that someone still must provide and fund that underlying connectivity.