What/how would it be destroyed? Some of the big relays might struggle, but that would only strengthen the network as more relays step up to help handle the load.
Discussion
Mature software like Mastodon, built four scale could not even handle it when Elon bought Twitter and they flex there. They had to upgrade their hardware.
Did you ever see Twitters Infrastructure? Its insane how many databases they have all repliacting and communicating together. Thousands of servers
If you ever run a nostr relay, you would quickly realize how far we have to go. Now its easy for anyone to host a relay but at scale ,you would need a private relay or limited relay so you wont get overwhelmed.
The database now of a relay grows way too much. Also, spammers can write to your relay as much as they want and theres nothing you can do.
I literally built relay software and run a decently sized public relay, so I am aware of the impact it would have. Especially on the memory hungry relay implementations that use pgsql.
As for spammers, that's inevitable. What matters is how you rate-limit the number of events they send.
And it would also mean paid relays being a more attractive solution. In fact, I wouldn't doubt if we eventually see nostr:npub1xtscya34g58tk0z605fvr788k263gsu6cy9x0mhnm87echrgufzsevkk5s decide to turn the Damus relay into only being accessible by Purple subscribers in the future.
Most free relays won't be sustainable, without hard capping how many events they choose to handle, or if there's a non-profit willing to keep things operational.
>As for spammers, that's inevitable. What matters is how you rate-
limit the number of events they send.
Yes but no. Every alternative to Nostr gives users Administrative controls. Nostr relays are basically a public notepad file that any person can write to as much as they want.
> And it would also mean paid relays being a more attractive solution
Paid relays are not a solution at all. Basically, Nostr will be so terrible that people will need to pay to have a good experience? That will hurt adoption.
If most free relays won't be sustainable, why would anyone choose Nostr over anything else? The Fediverse is thriving and growing and interacting across thousands of servers each day. Less Nostr relays = more centralization.
I didn't even get into the centralization of media uploads on Nostr yet and how the costs will sky rocket. That's another discussion.....kinda
> The Fediverse is thriving and growing and interacting across thousands of servers each day.
But you said earlier they had a scaling issue. I know as I was there for it too. Same issue with Lemmy.
There are also instance Admins who accept donations or charge a small fee. That's the equivalent of paid Nostr relay.
We'll see where things lead. Only time will tell π€
My point is that the relay software needs to improve now rather than later. Fedi can scale better than nostr now, their software is not as IO heavy either. Someone could even run a decent fediverse server on a Raspberry Pi with an SSD.
Absolutely, I think those refinements will come in time. AP has had a couple extra years to mature and with a lot more dev assistance. From what I've read, Nostr contributors still need to workout how to properly handle NIPs and a consensus on changes that doesn't break previous implementations. I'm hopeful, but not naive. I know this will take a long time to achieve any level of "overnight success". We'll learn from our mistakes along the way, and hopefully learn from others (AP, ATproto, etc) on what not to do.
What I'm proposing is almost already fundamentally here. For example, nostr:npub16c0nh3dnadzqpm76uctf5hqhe2lny344zsmpm6feee9p5rdxaa9q586nvr Primal already issues a primal.net lnaddress for everyone who uses the wallet. Now, all they have to do is use that same address as a NIP-05 too. Let the user pick their username at time of sign up, and that's all the user really knows about the key pair as it is only mentioned in passing to save nsec, placing the emphasis on the NIP-05 Nostr address. That's what they'd share to people and what others could use to find them. A more friendly UI/UX approach imo.
They require you to give a real name, email, and Telephone number.....I thought Coiners want anonymity?
When I tried the Primal app I used all fake details π Fuck any kind of KYC.