[rob74]

> The complexities of managing infrastructure, buying new hardware, upgrading software, replacing failed disks — had been abstracted away

...and replaced by the complexities of AWS. I mean, even in the pre-AWS days, it's not like you had to buy the hardware or replace the failed disks yourself, web hosters did that for you.

[Aurornis]

> it's not like you had to buy the hardware or replace the failed disks yourself, web hosters did that for you.

That’s an oversimplification. Yes, you didn’t go into the data center and replace the disk yourself, but it was a very different process than what you get with AWS.

There are still plenty of web hosts that operate the old fashioned way if people want to use them. There is a reason people prefer services like AWS though.

[ceejayoz]

Sure, but it often required a) you to detect and diagnose it on your own and b) a couple hours to a couple of weeks for them to agree and swap it out. Versus the ability to click a few buttons and have a brand new server on separate hardware.

[dangus]

AWS doesn’t have to be complex if those aren’t your requirements, you can use something like Lightsail. If all you need is a VPS, get a VPS. (Though if that’s what you need, I personally like Hetzner more).

But AWS needs to be complex to handle the needs of huge organizations.

I will also point out that the AWS console EC2 launching interface has come such a long way. So much to the complexity is handled for you.

[ignoramous]

> I mean, even in the pre-AWS days, it's not like you had to buy the hardware or replace the failed disks yourself, web hosters did that for you.

Given no single web hoster has grown to the size of AWS / Azure, it is safe to assume just which complexities the industry is willing to tolerate. I mean, Oracle still rakes in billions, despite everything.