Comment on ‘There isn’t really another choice:’ Signal chief explains why the encrypted messenger relies on AWS
KairuByte@lemmy.dbzer0.com 3 days agoI get your point, but that comes with a whole host of other problems. Take a look at Lemmy for instance, decentralized, yes. But also prone to problems stemming from that same decentralization.
who@feddit.org 3 days ago
Exactly what problems do you have in mind?
There is no reason to assume that a distributed incarnation of Signal would have the same design as ActivityPub or Lemmy.
KairuByte@lemmy.dbzer0.com 2 days ago
Massive lag coming from larger instances, instance moves or domain name loss causing the death of an instance, misconfigurations in general since those cause a plethora of problems.
who@feddit.org 2 days ago
Problems like those are unavoidable even on today’s Signal, because the service depends on internet peering relationships, internet service providers, mobile network operators, cell service, etc. Oh, and Amazon.
You usually don’t notice them because when any of those components experiences problems too often, affected users tend to get annoyed and switch to a more reliable one. (Also because you don’t expect to receive messages from as many people or as often as you do on Lemmy, so short outages are less likely to affect you at all.)
That would still be true in a distributed Signal, except that users could switch away from Amazon as well. Meanwhile, everyone not using Amazon would still be chatting during such an outage.
1984@lemmy.today 2 days ago
Thats true but what examples is there of successful distributed apps with tons of users?
ethicallysliced@lemmy.zip 2 days ago
Email
who@feddit.org 2 days ago
We’re not talking about a distributed app, but a distributed service.
Email.
The web.
The entire internet.
The postal service.