| 31 Jul 2021 |
 mark22k | * The others cannot be children if I do not accept incoming connections. | 16:59:46 |
 Kimapr | In reply to @mark22k:tchncs.de
The others cannot be children if I do not accept incoming connections. there's no difference between an outgoing or incoming peering | 17:04:32 |
| Kimapr | In reply to @mark22k:tchncs.de
The website says that every node has one parent. But I connect with several peers. How can that be? iirc the closest-to-root peer is chosen as your node's parent | 17:05:55 |
 Aminda Fantomino de Rozo | and the tree (with parents and children) is just the worst case scenario, in reality your traffic will take any shortcut they can if your peers (and their peers) are closer to the destination | 17:16:42 |
|  Мамба left the room. | 17:19:49 |
|  aagainb joined the room. | 17:35:59 |
|  else_ joined the room. | 19:26:18 |
| else_ left the room. | 19:28:01 |
| else_ joined the room. | 19:28:54 |
 tomz | In reply to @mark22k:tchncs.de
The website says that every node has one parent. But I connect with several peers. How can that be? This is the genius of yggdrasil.
Yes, every node can have many peers, the true mesh-networking style.
But for routing your messages it is relevant to pick the fastest one (to the root in this case) in order to produce a way in which each node knows how to find other nodes.
This means that if you have many peers, your node still selects exactly one to be a parent. That can change a couple hours later, so nothing permanent, but relevant for routing.
| 19:31:31 |
| tomz | check out http://[202:d7cd:bd04:6bbc:acc6:b002:da12:86c7]/ | 19:32:03 |
| tomz | or, as that one doesn't seem to work right now; http://51.15.204.214/ | 19:34:31 |
| Kimapr | In reply to @tomz:matrix.org
or, as that one doesn't seem to work right now; http://51.15.204.214/ that one doesn't work either :( | 20:07:35 |
| Kimapr | hmmmm | 20:08:10 |
| Kimapr | [kimapr@kdp_portable ~]$ curl https://51.15.204.214/
curl: (7) Failed to connect to 51.15.204.214 port 443 after 91 ms: Connection refused
[kimapr@kdp_portable ~]$ curl http://51.15.204.214/
curl: (7) Failed to connect to 51.15.204.214 port 80 after 0 ms: Connection refused
| 20:08:15 |
| Kimapr | that's some weird fuckery | 20:08:21 |
| Kimapr | >0ms ping with my internyet | 20:09:14 |
| Kimapr | >different latency on different ports | 20:11:08 |
| Kimapr | someone's messing with my localhost | 20:12:01 |
| Kimapr | as i have 1-5ms of latency to the router | 20:12:21 |
| Kimapr | it works correctly on tor browser | 20:14:36 |
| Kimapr | uuhh | 20:18:06 |
| else_ left the room. | 20:20:57 |
| else_ joined the room. | 20:41:10 |
 errhammr | 🤔 I've got an idea. Nested yggdrasil – yggdrasil over yggdrasil over yggdrasil... Dunno what this would be good for but is it possible? | 22:52:49 |
 Revertron | In reply to @errhammr:matrix.org
🤔 I've got an idea. Nested yggdrasil – yggdrasil over yggdrasil over yggdrasil... Dunno what this would be good for but is it possible? Parallel networks without connection in between. One outer, other inner :) | 22:54:32 |
| errhammr | Yes, something like that. So I'd need to disable multicast so that the outer and inner nodes don't see each other. | 22:55:36 |
| errhammr | Or maybe I want the inner net to communicate with the outer net. | 22:56:09 |
| Revertron | In reply to @errhammr:matrix.org
Or maybe I want the inner net to communicate with the outer net. This is the bad idea. | 22:56:37 |
| Kimapr | oh no | 22:58:36 |
