Unlike others, I’d expect a signicifant decline of posts as not only many users but also loads of communities would be lost. That’s why from my perspective users and communities should be evenly distributed across instances.
On top of that, there should be a feature to move entire accounts or communities to other instances. That way a community including all its members could just be migrated before a major shutdown.
Similarly, I think it would be a huge disturbance for the email system and possibly the entire internet world-wide if Gmail went down next month even though there are in theory plenty of alternative providers. Or supermarkets. If the IT of Walmart, Visa/MasterCard, Amazon AWS, Microsoft etc. have an outage it always has huge impact.
Lemmy as a whole isn’t that big and far from being critical infrastructure but we all want it to grow we should bear in mind that huge central services are always more risky than many small services.
That’s an amazing idea from my perspective which would really help the fediverse in general. Your proposal would solve not only the issue of instances disappearing completely but also temporary outages. For example Feddit.de as one of the biggest German instances had several technical issues lately. And as instances are typically administered by volunteers during their freetime that’s totally fine, accepted and expected. Such a failover concept would heavily reduce the pressure on their shoulders!
I’ve mentioned it elsewhere but the way Lemmy was implemented doesn’t make sense from a user perspective.
The hosting should be decentralized, there should only be one website.
Server admins would choose if they agree to host NSFW content or not and would have the power to stop hosting content from certain communities but it should have no impact on the user’s side.
All content should be hosted on three servers at all times so if one shuts down there’s still a main and a backup server until they’ve finished uploading the content to a third server. All the same content wouldn’t necessarily be hosted on the same servers, what’s important would just be that there’s triple redundancy.
So yeah, make it like any other website, just decentralize the hosting itself… and that’s exactly what providers for major websites do as a matter of fact, so just do the same thing but using servers owned by a bunch of people instead of one company that has server farms in multiple locations.
If you just decentralize only the physical hosting part how would you handle responsibilities such as moderation and other key decisions? If there’s one central instance deciding on what to allow and what to block or on topics such as advertising, trackers etc., wouldn’t Lemmy end up with similar issues as Reddit and other traditional social media websites?
Moderation would work the same way it does right now, admins aren’t moderators.
An admin doesn’t want to host content from community X? They just block it from their server, it doesn’t matter, it’s hosted on two other servers and they’ll take care of sending the data to a third one.
Donations would be easy to automate via crypto, your public key would be part of the code, donation goes to the community and is split between the hosting servers.
In addition to features for migrating communities and accounts I think the ability to set up a special type of instance that just archives everything from all the instances it’s federated to would be a huge benefit.
Something along the lines of an instance that doesn’t allow new content to be created, only consumed. This way if an instance were to permanently close we could migrate it’s communities to other instances from the archive.
This could also extend to migration of lost accounts, though ensuring the original account holder is the one making a request could become a nightmare of an overhead. The situation could be improved though if lemmy got some sort of feature for linking accounts across multiple instances.
Unlike others, I’d expect a signicifant decline of posts as not only many users but also loads of communities would be lost. That’s why from my perspective users and communities should be evenly distributed across instances.
On top of that, there should be a feature to move entire accounts or communities to other instances. That way a community including all its members could just be migrated before a major shutdown.
Similarly, I think it would be a huge disturbance for the email system and possibly the entire internet world-wide if Gmail went down next month even though there are in theory plenty of alternative providers. Or supermarkets. If the IT of Walmart, Visa/MasterCard, Amazon AWS, Microsoft etc. have an outage it always has huge impact.
Lemmy as a whole isn’t that big and far from being critical infrastructure but we all want it to grow we should bear in mind that huge central services are always more risky than many small services.
It’d be cool if you could specify a backup mirror community, so there could be an immediate failsafe.
That’s an amazing idea from my perspective which would really help the fediverse in general. Your proposal would solve not only the issue of instances disappearing completely but also temporary outages. For example Feddit.de as one of the biggest German instances had several technical issues lately. And as instances are typically administered by volunteers during their freetime that’s totally fine, accepted and expected. Such a failover concept would heavily reduce the pressure on their shoulders!
Gmail getting the axe would be so bad it would get a section in history books
I’ve mentioned it elsewhere but the way Lemmy was implemented doesn’t make sense from a user perspective.
The hosting should be decentralized, there should only be one website.
Server admins would choose if they agree to host NSFW content or not and would have the power to stop hosting content from certain communities but it should have no impact on the user’s side.
All content should be hosted on three servers at all times so if one shuts down there’s still a main and a backup server until they’ve finished uploading the content to a third server. All the same content wouldn’t necessarily be hosted on the same servers, what’s important would just be that there’s triple redundancy.
So yeah, make it like any other website, just decentralize the hosting itself… and that’s exactly what providers for major websites do as a matter of fact, so just do the same thing but using servers owned by a bunch of people instead of one company that has server farms in multiple locations.
If you just decentralize only the physical hosting part how would you handle responsibilities such as moderation and other key decisions? If there’s one central instance deciding on what to allow and what to block or on topics such as advertising, trackers etc., wouldn’t Lemmy end up with similar issues as Reddit and other traditional social media websites?
It doesn’t require a central authority.
Moderation would work the same way it does right now, admins aren’t moderators.
An admin doesn’t want to host content from community X? They just block it from their server, it doesn’t matter, it’s hosted on two other servers and they’ll take care of sending the data to a third one.
Donations would be easy to automate via crypto, your public key would be part of the code, donation goes to the community and is split between the hosting servers.
In addition to features for migrating communities and accounts I think the ability to set up a special type of instance that just archives everything from all the instances it’s federated to would be a huge benefit.
Something along the lines of an instance that doesn’t allow new content to be created, only consumed. This way if an instance were to permanently close we could migrate it’s communities to other instances from the archive.
This could also extend to migration of lost accounts, though ensuring the original account holder is the one making a request could become a nightmare of an overhead. The situation could be improved though if lemmy got some sort of feature for linking accounts across multiple instances.