Signal was just one of many services brought down by the AWS outage.
Yes, but you can have redundancy though. Obviously it comes with a cost, and I don’t know if Signal can afford it.
Submitted 5 hours ago by BrikoX@lemmy.zip to technology@lemmy.zip
https://www.theverge.com/news/807147/signal-aws-outage-meredith-whittaker
Signal was just one of many services brought down by the AWS outage.
Yes, but you can have redundancy though. Obviously it comes with a cost, and I don’t know if Signal can afford it.
They can’t afford what they have now
That was… enlightening. I can’t imagine the scaling they had to do from day 1 to now.
It is true that there really isn’t another cloud provider that they could choose. All of the other cloud providers (major and minor players) are prone to the same sort of systemic failure. But it isn’t true that they didn’t have another choice.
The solution to service failure is redundancy. Making the redundancy as different as possible makes it even more resilient. In this case, that would be having redundant servers on other cloud providers which can be used in the event that the main one fails. Even better if they can use all of them simultaneously to share the load and let failover happen more gracefully.
That is very pricey
Right, OK, but Signal sustains itself on charity.
I don’t think that’s necessarily incompatible with what I suggested. They could just leave the backup servers offline until they’re actually needed which shouldn’t cost them anything (or at least not much; some cloud providers charge for a VM’s storage usage regardless).
Assuming that Signal’s servers were designed by competent engineers, the engineering cost to make a change like this shouldn’t be that bad. Though judging by Whittaker’s comments, that may be a bad assumption.
Signal was down? Didn’t even noticed
who@feddit.org 2 hours ago
To me, this reads like sophistry.
What happened here is a predictable result of Signal’s design. They chose to build a centralized messaging system. This made things significantly easier for them than a distributed design would have been, but it has its drawbacks. Having single point of failure is one of them. (In this case, that single point is Amazon.)
Trying to direct the public’s focus onto cloud providers instead of acknowledging this fundamental shortcoming in their design is, frankly, disingenuous. Especially coming from someone in Whittaker’s position.
While we’re at it, let’s also acknowledge that centralized design in messaging networks are problematic not just because of (un)reliability, as seen here. It’s also a single point of attack for any entity seeking to restrict, shut down, or track people’s communications with each other. End-to-end encryption cannot solve those problems.
possiblylinux127@lemmy.zip 1 hour ago
Signal is user friendly and reliable
While I don’t agree with some of their choices they do have a point here.