Ok so maybe I’ve missed something over the years. I know there are subnets for specific things etc. but what about a new type of subnet, that anybody can buy the correct hardware and run nodes from their basement/bathroom/car/datacenter/bare metal and if there’s 3+ of them it makes some sort of community subnet which is still on mainnet and thus icp will win.
Can someone actually explain to me why we cannot have permissionless nodes. I want a real, easy answer. Not some bs technical jargon.
If the nodes don’t get rewards, I don’t see a reason why not. If they would, then it opens up discussion in my opinion.
Some questions: Who will have authority on what nodes join? When are they swapped out? How will you evaluate this.
A counter argument I can think of: a subnet can influence the availability of other subnets. So I think the Cloud Engine format would suit this use-case better. As they don’t have unbounded await calls/cycle management.
Ok so I am not smart, but I’m pretty sure on ethereum chain (some old Dino token idk) there is a minimum amount of eth tokens you must stake in order to validate the network. This in turn prevents garbage validator or nodes because of the cap requirements to entrance and what I think the kids call slashing?
I am not certain.
maybe a community node subnet dispenses rewards at a discount against “normal” nodes?
Wait, so yes in theory cloud engines is correct but do those engines connect to mainnet icp? Or are they just their own cloud ie if I deploy a canister on my personal cloud net, can I interact with a cansiter on mainnet without latency or other unknown issues?
Yes, CEs can talk to mainnet. They get somewhat less guarantees than normal subnets (i.e. unbounded vs bounded calls). So you’re canister implementation will need to take this into account. But aside from that all the same.
Oh, also except cycles, you can not attach any cycles to calls… which depending on the use-case might be blocking. This can be solved with a cycle-attaching-proxy-canister tho.