Or have they stopped needing that (I haven’t followed it for quite a while)
I want to start joining, but I’d like to run the client in a virtual machine with zero configuration.
Let me know what’s still needed, thanks guys.
1 Like
Port forwarding is still needed behind a NAT router - to run a node but not the client.
So anyone can try the network out for file upload, download, token transfers etc.
You can also run a local testnet if you want to learn about running nodes on your own.
3 Likes
Clients work just like that, but for the node you need port forwarding.
Nice thanks guys, I will definitely at least mess with the client in this next testnet.
Don’t know if I’ll bother with the node, maybe or maybe I’ll wait a few months as we get closer to launch.
4 Likes
Does anyone here know how the bitcoin node software works straight out of the box, 
% decentralized. can’t we use the same trick?
1 Like
There is a big difference; for bitcoin it is enough for a node to connect to n clients (actually, 1 is enough for functionality, but not for good security) and it does not matter if all are outgoing. For safenet, a node must be reacable for every client on the network. Cjdns have a similar problem, and a solution too. Not sure if it would be applicable to safenet…
Please correct, if I’m wrong…
2 Likes
BTC full nodes have used uPnP for a while, if it works great, but not everyone can run a bitcoin full node. The software will still run, if port forward failed, but in client mode.
4 Likes
Will running at least one node (without NAT configuration) along with the client be the standard setup on network launch for “normal users”?
I have doubts (I actually ran some simulations) that the network economics are sustainable if nodes are only run by farmers.
Firstly depends on what you call farmers. I, and I think many here, call farmers anyone who runs nodes. One or many nodes.
I gather you mean datacentres and VMs in the “cloud” type farmers. And I would agree with you that its not good at all if that was the case. For one it jacks up the basic running costs and centralises control too much.
In any case people are successfully running nodes behind a NAT using port forwarding. Many gamers already know how to port forward because their games required it (uPnP or manual setting). Anyone running a local game server also needs to know.
The library used is enabling autonat (I think that the name) and currently debugging it so that is expected soon which will allow for nodes behind NAT without any user effort needed.
People have reported that using a VPN allowed them to run nodes on their computers even though they are behind a NAT since the VPN is doing the work anyhow. Remember that if the ISP doesn’t know what your traffic is then so too the VPN provider cannot know either. In AU its the law that ISPs records all meta data, so it makes no difference if the VPN provider claims no logs but keeps warehouses of disks with logs or not.
2 Likes
TL;DR:
My view is: If we allow storage being provided by commercially driven entities (like bitcoin mining), the network growth will stall as storage prices will rise too high.
My definition of farmers is: Someone running nodes to gain profit from rewards instead of primarily using the network to store data. My assumption is that the farmers by this definition are driven by profit (this assumption is probably true if you follow the discussion in the farming hardware thread) just like bitcoin miners. In my simulations of the network, unfortunately there are several stall scenarios where network growth comes to a halt.
Farmers likely only provide storage as long as the profit is higher than their cost (hardware, power, maintenance) - not all of them will be idealistic. At the same time, users of the network will only store data if cost is not prohibitive - not all of them will be idealistic. As storage rewards depend on new data being stored (“pay once”), my simulation runs into a complete network stall (relatively full network → storage price rising → farmers join, but at the same time users store less and less → less reward for farmers → farmers leave → no node growth).
The storage token distribution for existing data should counteract this effect somewhat - especially as its high initially.
If I assume every user also runs a node, the network grows quite nicely and storage cost keep being low. This is why I think “normal users” running nodes and providing storage is absolutely crutial. But then, it needs to be “dummy proof” to run a node.
This should also be in line with the original vision of the network running on everyday devices.
2 Likes
Yes we need a set and forget style of node install/run. And we’ve been told nodes behind NATs definitely should be possible.
With your definition of farmer, to me, seems contrary to what most here say a farmer is. Farmer is one who runs a node (at home or commercially) and has been used less and less as the network develops as it seems to not quite fit. Some say node operator now, or node runner. But we all know farmer is a node operator and you differ to say its commercial running.
Perhaps I can suggest to save any confusion, that you say something like farm for profit or better yet commercial farmer. This is solely to save confusion when others read your posts. Just a suggestion.
Might I suggest that your simulation might be missing a factor and that while its true that as price to store rises less will store, it also holds that as farmers join and price drops then more will store thus farmers still earn reasonable amounts. Remember nodes gain chunks as well as shed chunks. Thus when network is 70% full most nodes are around 70% full and as fullness drops due to new nodes joining to say 50% then most existing nodes drop to around 50% full and the new nodes take on churned chunks to be around 50%