4 ms·
Peer Name Resolution Protocol (2009)
- dfischer 6y agoThis was a bit of a wow for me. Good concepts here that are just now being discussed again. A way to handle distributed routing and name resolution. Projects like Yggrasil and CJDNS are similar. Some nice ideas here.
- WorldMaker 6y agoI made the mistake of trying to build a project using PNRP at one point (via WCF's terrible aborted peer-to-peer stack). PNRP had a lot of great ideas, but the actual Windows implementation was sadly temperamental (and debugging it a bear when it wasn't working correctly, which seemed to be every time it was newly installed and also half the time you tried to use WCF's peer-to-peer stack). In a theoretical standpoint it's a shame that mDNS (fka Bonjour) was the standard that won over PNRP, but in a practical "unix philosophy worse is better" standpoint it was easy to see why mDNS won. (And why it's taken so long for projects like Yggdrasil and CJDNS and IPNS to return to some of the PNRP ideas years later, as mDNS was "good enough" for the easy immediate need, versus the more long-term goals of PNRP.)
- dfischer 6y agoExactly why I was wow’d by it: my recent research into p2p systems that could replace BGP and falling upon Yggdrasil. Thanks for the anecdotes!
- aeternum 6y agoThis design was based on Chord, one of the first distributed hash table protocols: https://en.wikipedia.org/wiki/Chord_(peer-to-peer) https://en.wikipedia.org/wiki/Chord_(peer-to-peer)