|
|
|
|
|
|
by waheoo
2122 days ago
|
|
|
I don't really know why we need to discuss this now but the point has nothing to do with this. Small sites don't really need that decentralization, sure it's nice that it's easy but this isn't the problem. The problem is that in aggregate you end up with an internet with a single point of failure. And even if as it is now isn't a big problem, who's to say that one day it is a huge problem? |
|
|
And the other problem, which I'm pointing at, is that without using something like Cloudflare, you end giving yourself more points of failure (more pressure, DDoS, lack of easy load balancing, most costs and devops to implement those yourself, etc).
And each site doesn't care if 10000 others go down together -- if anything that's good, if their competitors go down for a while. They care for their own status...
>And even if as it is now isn't a big problem, who's to say that one day it is a huge problem?
I'd say periodic mass failures should inform our usage and dependance patterns of the internet so that we're not 100% dependent on it 24/7, in which case sites going down together never becomes "a huge problem".
In other words, one way to never have it be a huge problem is to make the internet perfectly decentralized (which is impossible anyway -- first because sites people care about is a power law distribution, e.g. Google, MS, Amazon, stores, app stores, etc, so if Google goes down there's a disruption to billions of people, even if millions of lesser sites are up that much fewer care about, -- and second because critical instrastructure is shared, e.g. undersea cables etc.
The other way to never have it be a huge problem is to learn and adapt to situations when sites might be done, and build resilient alternative ways of operation (analogue, if need be).