Was thinking the same thing
Let's hope more people deploy single-click relay services. :)
Then why have private relays?
- higher cost per user
- economies of scale are thrown out the window
- a larger mess to maintain
- more connections for clients
benefits
- your relay is separated for no fucking reason
Self-custody is always better. Harder and costlier, but better.
it is not self custody if your relay is run by some random person with total control to kill it at any time
just because there are multiple relay endpoints doesn't mean it is not a single entity and therefore more decentralized
read:
View quoted note βit is like saying some cloud hosted ln node provider like voltage which holds your keys is self custody. it is not. you just have access to the node
Big fan of Amethyst, but this isn't the right move imo. What happens if there are uptime issues with the service? Not to mention the centralisation as already mentioned. Greater emphasis should be placed on default relays that are segregated and known to be reliable. Even though I agree completely with the mission of relaying.io to put ownership in hands of individuals. Also, what infra is relaying.io running on? Would appear AWS... Oof
Then let's make sure people can use their
@Umbrel βοΈ effectively.
NIP 65 is not only for relaying.io. It's a great way to disperse all notes into millions of relays.
My thoughts exactly
And you are ok with being ruled by 5-10 large relays? Your data is their property. They can do whatever they want with your stuff.
Huge no, no to me.
not everyone can port forward and set up tls.
if you think that clients (especially mobile ones!) will be able to tolerate more than 100 connections (thousands if you follow a lot of people) then you are being blinded by the prospect of a new shiny thing you can claim your client supports
there will be a lot of relays. small and big ones.
but the idea of personal relays will not take off and will be a thing only certain people do.
What about Umbrel?
Do you want to see people actually using their Umbrel to host their data and allow other people to download info directly from the person's main repo?
Seriously, decentralization is better.
yes because I am not a fool that stores their data only on relays with no backup and it is very unlikely all those relays will attempt to censor me at the same time
You have no idea what mobile phones can do :)
It seems, imo, the current structure makes the most sense. Keep the current, reliable relays as defaults, but educate users on how to add their own relay as a failsafe.
if you pick your relays with enough diversity of the owning entities then you have a very low chance of being censored at the same time with no time to publish your own relay
personal relays will only be a tool highly censored people and some hobbyists use
You could use both, have βpersonalβ relays and some larger but not too large ones. Just a thought
it is neither good for the userβs mobile data, or their battery life to process that many connections
after some point you connect to enough relays that it starts to break other stuff than the app
Let the users what's good for them or not.
why have both personal and large relays?
personal relays are additional strain on peopleβs clients, and it is very rare you actually need it (if you are highly censored for example)
* decide..
there are a lot of real cases where you let people decide and they fall for deceptive marketing
Large relays give you reach, relays that you run yourself give security over you own data. I would only need to run one personal relay to secure my data and that would take me up to 8 at time of posting, not that many really
it's still their choice.
I could sell subdomains of my big relay that only show someoneβs events and claim it is a private relay. so what is the difference between βprivate relaysβ and big paid relays? the fact that you have a fucking separate domain
Clients should do integrated backup
Then do it, let's see if they go for it.
Most users (especially new ones) are utterly confused about relays.
Letting them troubleshoot this on their own is a huge hurdle.
βwe should allow deceptive advertising since people need to have a choiceβ
They can.
they CAN. but what is the probability of that happening?
a lot of stuff relies on βeffectively impossible to happenβ, not β100% impossible to happenβ
Perhaps, Iβm not smart enough to know if thatβs a good idea π but until they do we rely on relays
Basically stores all your events so you can migrate them
maybe to stop centralization we should stop shipping a default relay list for everyone snd have a larger pool + pick a random set for everyone
that is not what I said. if the only effective difference between private relays and big paid ones are a fucking subdomain, why waste resources on having so many connections and waste bandwidth?
The answers that noobs get for relay questions are:
1. Copypasta this big Acctβs relays
2. Pay for relays
3. Have to find the sweet spot - 7 to 14 relays
Not to mention if relaying.io EC2 instances are all in the same region, then RIP ping for some users.
Sounds good but you end up with a bunch of flaky relays that cause more trouble than they solve. Would need some sort of way to know if they are good reliable relays or not
This shit is confusing to me too and Iβve been here since Feb. Iβm not a dev, not super tech savvy. Iβm speaking for new users and adoption.
Actually, it's not that low. Any political instability in Canada can have unpleasant consequences. For some random reason, a significant number of main relays are located in this country.
But I agree, it's not worth getting too carried away here either. It's too easy to get paranoid.
I think there are a lot of relays in other locations too
Canada, really?! That would be highly ironic
If they can force banks to freeze accounts for the trucker protesters, imagine how easily they could shut down relays.
Generally, wouldn't rate them as competent in locating relays--unless they had help from Big Brother
People have options to use different clients if they want to :) no one is forcing anyone
NIP-65 wasn't written with personal relays in mind. Even if we all gather on just 30 relays (we have 200-800 or more already today), clients still need to know which ones have your events in order to follow you.
Each connection has TLS overhead, and that's probably the most expensive part for a mobile phone. If we all disperse into personal relays, following a thousand people on a personal phone at 500 personal relays would probably (as you say) be too much for a phone to handle (battery wise). Therefore, there is still room for relay services of various types.
This isn't an either/or. I think NIP-65 can work along side aggregating relays.
I have a personal relay. But I post to 7-8 relays. And most of those 7-8 relays are very popular ones. I expect smart clients will read my events from the popular relay that they are already connected to, and not bother connecting to my personal relay which at most could only gives them events from 1 more person.
The idea of the personal relay is to have an authoratitive source and a backup, not to be the content-distribution network.
Personal relays should then be hosted by the person which is very difficult β I see no real point of relay hosting services since it is more cost effective and efficient in general + they provide no additional censorship resistance guarantees than a regular big relay
I believe this is what
https://nostr1.com/ does.
