Reliability is not even how the cloud got sold to the C Suite. Good God, when my last company started putting things on Azure back in 2015 stuff would break weekly, usually on Monday mornings.

No, the value proposition was always about saving money, turning CapEx into OpEx. Direct quote from my former CEO maybe 9 years ago: We are getting out of the business of buying servers.

Cloud engineering involves architecting for unexpected events: retry patterns, availability zones, multi-region fail over, that sort of thing.

Now - does it all add up to cost savings? I could not tell you. I have seen some case studies, but I also have been around long enough to take those with a big grain of salt.

▲

mosura 7 hours ago | parent | next [-]

No the value was bypassing IT.

You no longer needed them to approve a new machine, you just spun it up how you want. Sped things up massively for a while.

▲

darkwater 7 hours ago | parent | next [-]

That might have been true for some kind of organization, but definitely not for every kind. On the other side, there were start-ups that wanted the elasticity and no commitments. But both sides at least partially liked the "it's not on me anymore" feature.

▲

ocdtrekkie 7 hours ago | parent | prev [-]

It's amazing how there's so many cybersecurity incidents now. Bypassing IT will always backfire spectacularly, IT is the people that stop you from dumbing.

▲

mosura 7 hours ago | parent [-]

The opposite was/is true. If your cloud box can only be used by two people and IT don’t even know about it then IT can never be persuaded to provide the keys to the rest of the company as they were predisposed to doing.

I saw this stuff too many times, and it is precisely why the cloud exploded in use in about 2010.

▲

ocdtrekkie 6 hours ago | parent [-]

What you're telling me is two people potentially have regulated or confidential data not secured by IT, which nobody knows if got leaked.

For many organizations, that's literally illegal, and anyone who does this should be fired.

▲

mosura 6 hours ago | parent | next [-]

One notable example was signing keys for builds for distribution actually. And IT had a habit of handing them out to absolutely everyone. Being able to audit who did the signing was done in spite of IT who could, of course, never be persuaded of the merit of any process they don’t own.

But sure jump to more conclusions if you want.

▲

ocdtrekkie 5 hours ago | parent [-]

I won't discount your IT can be bad, but also if you're keeping something as core to your security as signing keys somewhere your IT can't audit, you are just as bad. And your IT won't be the ones fired when your keys leak.

	▲	mosura 3 hours ago \| parent [-]
		You are under the erroneous impression IT would be fired for leaking keys and not simply impose a new process that blames everyone else. And this is in Fortune 500 of course.

▲

pixl97 6 hours ago | parent | prev [-]

>, that's literally illegal, and anyone who does this should be fired.

I mean yea, but who knows how long that box would sit around before it was discovered.

▲

dig1 7 hours ago | parent | prev [-]

> ...does it all add up to cost savings?

IMHO it adds, but only if you are big enough. Netflix level. At that level, you go and dine with Bezos and negotiate a massive discount. For anyone else, I’d genuinely love to see the numbers that prove otherwise.

> There's zero personal responsibility

Unfortunately, this seems to be the unspoken mantra of modern IT management. Nobody wants to be directly accountable for anything, yet everyone wants to have their fingerprints on everything. A paradox of collaboration without ownership.

	▲	SJC_Hacker 6 hours ago \| parent \| next [-]
		> IMHO it adds, but only if you are big enough. Netflix level. At that level, you go and dine with Bezos and negotiate a massive discount. For anyone else, I’d genuinely love to see the numbers that prove otherwise. It adds if you're smart about using resources efficiently, at any level. And engineer the system to spin up / spin down as customers dictate. For situations where resources are allocated but are only being utilized a low percentage (even < 50% in some cases), it is not cost effective. All that compute / RAM / disk / network etc. is just sitting there wasted.
	▲	RajT88 7 hours ago \| parent \| prev [-]
		Cloud providers have formalized these deals actually. If you promise to spend X amount over Y period, you get Z discounts. And this is not reserved instances, this is an org level pricing deal. Some have been calling it anti-competitive and saying the regulators need to look at the practice.