That sounds pretty amazing. Excellent work and thanks for contributing, I hope this gets implemented. On 16 Oct 2015 11:02, "Рысь" <[email protected]> wrote:
> Hello dear Tincers! > > I recently developed an extension to tinc 1.0.x protocol which > introduces automatic and decentralized hosts update subsystem. > > The idea is to provide stable protocol extension to tinc which will do > all the dirty work of spreading information about new hosts in network > across all nodes by powers of tinc itself. > > If you're interested, you can take a look at the diff made for tinc > 1.0.16 here: > > https://github.com/siblynx/tinc-1.0.16_hostupd/commit/6a6cc34d6d80696ea525956375de59c7f752fa42 > > The code introduces two request types, and uses RSA signatures to > prevent tampering. It also introduces some sort of privileges: who can > and who cannot spread updates. Each node which has child connections > forwards update requests. > > Users can decide to whom to trust, and can turn off updates for their > own node. They also can setup completely trustful network in which > everyone will be permitted to update the whole network. > > Please see source file src/protocol_hostsupdate.c, it contains more > information about the extension implementation and how to configure > each node for it. In future I probably will move it to a README.hostupd > file. > > This is a motivation from the lesson I learned recently when my central > update service crashed inside my network with huge disk failures, and > that node's key became unrecoverable (yeah, I lost it, no backups, it > was a cheap "PC" router), resulting in unmanaged network for a few > weeks. It served updates centrally via http, and it's URL was the only > one which was recorded in update scripts and binaries for windows > everywhere. > > Even when I started using tinc, I seen that it misses something > important when I was forced to update hosts by hand for some time. > > This work is still experimental, I even test it now, and if you will > want to merge it into current or future versions of Tinc, you will > probably need to trim it of my extensions for my network. It probably > _contains_ bugs I still did not found, and I am not a pro in > cryptography (albeit having some experience with in past). > > This work is currently (my own code) is licensed under GPLv2 only. I > will happily unlicense it (making it public domain) if it will go into > any versions of Tinc, just with mentioning my contribution in THANKS > file. If not, it will stay as is. > > With best wishes! > > P.S. I found Tinc a great code inside when I developed extension for > it. I was able to quickly understand configuration and requests > subsystem, and the rest was no a problem. It's a shame that Tinc still > gets little attention in VPN area! > > -- > http://lynxlynx.tk/ > Power electronics made simple > Unix and simple KISS C code > _______________________________________________ > tinc mailing list > [email protected] > http://www.tinc-vpn.org/cgi-bin/mailman/listinfo/tinc >
_______________________________________________ tinc mailing list [email protected] http://www.tinc-vpn.org/cgi-bin/mailman/listinfo/tinc
