A few days ago Amazon Web Services went down for a bit and the entire Internet experienced a pretty sizable disruption. Why did this happen? Well AWS is a very popular service. The uptime is decent. It's cheap. It's mostly reliable. It's a centralized cheap reliable product on the cloud that a lot of people use... so when it does end up going down for a bit it affects every service that uses it, which is quite a lot.
AWS runs thousands of servers, so a disruption in AWS intrinsically implies a similar disruption across dozens if not hundreds or even thousands of applications. Of course in crypto this forces us to engage in an extreme reality check. When AWS goes down and a lot of crypto stuff goes down as well, we all get reminded just how interconnected a decentralized system can be to a centralized one. After all: if "decentralized" systems collapse when a single provider goes down how decentralized is it really?
If we take a look at that visual representation of what "decentralized" actually means it ends up being highly subjective and open to interpretation. If we imagine AWS being one of those hubs that connects to several other nodes... then yes, the system is still completely decentralized even if many people use AWS.
https://x.com/edict3d/status/1980343842382754000
People use AWS because it is convenient, but that doesn't make it a reliance issue. Notice how the timeline isn't full of people talking about how they'll never use AWS again and this outage is unacceptable in every way. That's because AWS has a very good track record and making that claim would appear like a childish response.
It's not hard to calculate just how long AWS has been down over the last 4 or 5 years... or even the past decade. The amount of time we talking about here is hours. Hours of downtime vs years of uptime. Pretty good. The standard for services like this has always been 99.9% uptime or better, and they are easily achieving that number.
So then why when AWS goes down are all these people running around larping about how nothing is decentralized? Can they not be bothered to wait an hour to use their favorite app? I tried to log into Venmo during the crash and it threw and error. No big deal, I'll just try again later. How much of an inconvenience was this crash really in the grand scheme of things?
The answer to this is: basically zero. Nobody that uses AWS is going to stop using it because once every couple years it goes down for a few hours. Talk to me again if this becomes more of a problem or they start trying to censor or otherwise mess with crypto nodes. Obviously if you run a node on AWS and it stops working as intended... you're going to stop using AWS. This is not an option for many other centralized WEB2 services that absolutely depend on AWS.
There are some companies that have opted to put all their infrastructure up on the cloud, and the amount of work and money it would take to migrate to another model is completely unviable. Crypto does not suffer from this issue. A crypto node is a single server that can be moved to a different server at a moment's notice. Even if a blockchain went down because AWS went down they would still have the power to act and quickly correct the situation.
I don't actually know what the guy in that tweet is bellyaching about but it looks to me like something on Ethereum. Maybe a browser extension wallet or website for some random altcoin or exchange. Yes, those things are much more centralized than the collective network itself. Nobody ever said these systems were distributed with a dozen backups for every little thing. That would be absurd and expensive to the point of complete prohibitive financial unviability.
nodeIf we are taking a look at the node I run as an example it's in a very similar situation. This thing is still running on Privex servers. Not even a dedicated server but a virtual machine sharing resources with other people. Pretty good chance at least one other node on Hive is running from the exact same hardware as I am.
Clearly, that is not the ideal scenario, and I've told myself on several occasions that I should make the switch to a better setup, but never seem to get around to figuring it all out. This doesn't change the fact that I would be forced to very quickly figure it out if the setup I currently have decided to stop working. As a general rule of thumb crypto has exponentially higher maneuverability when it comes to things like this than any kind of hulking WEB2 product. Oh yeah and congratulations to Hive for 100M blocks. What a crazy milestone that is.
Crypto bros who complain about collapse in tandem with AWS have lost the plot and forgotten what the point of this tech actually is. Did the AWS crash censor anyone's operations? Did the AWS crash print money out of thin air? Did the AWS crash create even a single meaningful adverse affect on crypto as a whole? No. It did not. All the larping we saw was just idealists complaining about how their completely unrealistic expectation/definition of "decentralization" was not achieved. Using a service does not imply reliance on that service within a modular dynamic ecosystem.
Of course this is no excuse to get complacent either. Just because people exaggerate these events for online attention doesn't mean the points being made are completely invalid. We should be striving to go our own way and build up fresh from a new foundation. In the meantime we need to accept that this is a reactive ecosystem that can respond quickly to negative stimuli. Nobody said the transition phase wouldn't be awkward.
