Admins Wonder If the Cloud Was Such a Good Idea After All (theregister.com) 119
After an initial euphoric rush to the cloud, admins are questioning the value and promise of the tech giant's services. The Register: According to a report published by UK cloud outfit Civo, more than a third of organizations surveyed reckoned that their move to the cloud had failed to live up to promises of cost-effectiveness. Over half reported a rise in their cloud bill. Although the survey, unsurprisingly, paints Civo in a flattering light, some of its figures may make uncomfortable reading for customers sold on the promises from hyperscalers. Like-for-like comparisons for a simple three-node cluster with 200 GB of persistent storage and a 5 TB data transfer showed prices going from $1,278.58 in 2022 to $1,458.68 in 2024 on Microsoft Azure.
For Google, the price went from $1,107.61 to $1,250.35. According to Civo's figures, the cost at AWS increased from $1,142.46 to $1,234.59. "The Kubernetes prices were taken from the hyperscalers' very own pricing calculators," a Civo spokesperson told The Register. In the IT world, there is an expectation that bang for buck increases as time goes by, but in this example, prices are rising faster than the rate of inflation, and what customers receive for their money remains unchanged.
For Google, the price went from $1,107.61 to $1,250.35. According to Civo's figures, the cost at AWS increased from $1,142.46 to $1,234.59. "The Kubernetes prices were taken from the hyperscalers' very own pricing calculators," a Civo spokesperson told The Register. In the IT world, there is an expectation that bang for buck increases as time goes by, but in this example, prices are rising faster than the rate of inflation, and what customers receive for their money remains unchanged.
and they vendor lock you in to make it hard to ge (Score:5, Insightful)
and they vendor lock you in to make it hard to get out.
Meaning there is no true free market happening... (Score:5, Interesting)
Think government should regulate this a bit. Push them to create a shared compatibility layer/API. Custom parts are possible if not covered by the shared spec already, though new entries should be added over time. Heck, maybe for AI stuff too (wonder if that's the final result for the NVIDIA + CUDA government worries?)
Force them to become commodities. Which creates true competition. Which lowers the costs of switching providers which creates incentive to compete on price/cost.
I've avoided looking much into cloud implementation stuff in part because I'd read that every place had their own version/flavor. And I'd hoped it would eventually migrate to a single method. When did cloud stuff start becoming a buzzword? Think it's been a while...
That's only true if you rely on their expertise. (Score:4, Interesting)
Re:That's only true if you rely on their expertise (Score:5, Insightful)
My experience is limited to AWS, but... It depends on what you are doing. If you are just using the cloud for pure compute, like EC2 or ECS/EKS, then yeah, there is very little lock in. But if you are using their more dynamic services like Elasticache, SQS/SNS, crypto services and whatnot, then you are generally writing to the AWS SDK, which does end up infecting your back-end code. Sure, you can wrap the AWS-specific stuff in a service interface that hides the fact you are talking to AWS, but you'll have to re-write that service implementation if you hop to a different cloud provider or an on-prem backend.
Re: (Score:2)
Yeeeeah! More APIs, designed by government officials encumbere with red tape. Definitely going to make it easier
Re:Meaning there is no true free market happening. (Score:4, Funny)
Re: Meaning there is no true free market happening (Score:3)
And require you to click a button ever time you make a call to âzprotectâoe you. The whole cookie confirmation dialog fiasco in the EU is exactly why this is a bad idea
Re: (Score:2)
But this takes more work and we all know how people feel about that...
Re:and they vendor lock you in to make it hard to (Score:5, Informative)
Re:and they vendor lock you in to make it hard to (Score:5, Insightful)
Managers are learning the same lessons they learned with outsourcing. Nobody takes care of you like you.
Re:and they vendor lock you in to make it hard to (Score:5, Insightful)
Pain being
cost
resolution on issues
wanting something special
feature requests
prioritization of work
support level knowledge and skill
It is a customer nightmare, and one any higher up would have trouble explaining to their bosses as to why things arent getting done. The biggest reason? "Well... we just arent big enough for the vendor... the business we pay, to care about us"
Re: (Score:3)
Sometimes you can't avoid SAAS. When I wrote a custom content delivery system for a new, large customer last year, their IT didn't want anything to do with the server being housed in their data center. On the other hand, they want to have all say in where the server was. It would have been cheaper for me to purchase two servers and co-locate them in data centers in 2 different cities to provide redundancy, but that wasn't good enough for them. They wanted AWS or Azure level services (redundancy, securit
ALL only partially correct (Score:3)
https://en.wikipedia.org/wiki/... [wikipedia.org]
Gartner 7 year technology hype cycle.
CIO/CEO/VP and other management have had more than 5 years for an on prem to cloud 'transformation' and all need a new
'miracle cure narrative' to sell to each other, investors, wall street, underlings in order to
- Justify their stock based compensation
- Justify their status in the 1% wealthiest club
- Create a springboard to go to the next higher paid, higher profile job and eventually 'retire' and live off corporate board money and stock
Re: (Score:2)
It depends on your circumstances. For a lot of SMEs the support they get from cloud providers is better than what they get from their staff. Is Dave the IT guy going to respond quickly on Sunday afternoon to have everything up again by Monday morning? Will the company shell out for redundant, unused infrastructure?
It's really the cost that is the issue. Cloud can be cheap if you are careful, but it can also be really expensive if you aren't.
Re:and they vendor lock you in to make it hard to (Score:5, Informative)
Just ask anyone who has been in IT for 40 years. The old timers knew were the Cloud was going. Spend lots of $ to move, then the companies extort you.
Happened in the 70s and 80s, happening now. Just a big circle.
Re:and they vendor lock you in to make it hard to (Score:4, Insightful)
Just ask anyone who has been in IT for 40 years. The old timers knew were the Cloud was going. Spend lots of $ to move, then the companies extort you.
Happened in the 70s and 80s, happening now. Just a big circle.
The only thing that puzzles me is why it hasn't swung the other way. It usually swings the other way.
Also, why companies haven't decided to move server infrastructure and data storage to on prem Linux?
I've heard lots of answers- which usually lead to the conclusion that people making decisions have no idea what they are doing.
Though, as I'm approaching 60, I have noticed that most of my junior engineers are really more clerks that don't know how anything really works.
The number of times since 2000 young people have looked at me like I discovered fire is uncountable.
Re:and they vendor lock you in to make it hard to (Score:5, Interesting)
The only thing that puzzles me is why it hasn't swung the other way. It usually swings the other way.
It is starting to but you're not going to see medium sized businesses (read: most businesses) make the switch because the upfront expense is not only unjustifiable... in most cases it is completely impractical.
If you're using AWS CDK and the proprietary cloud bullshit that comes with AWS, and the code was written by developers over years who were told that "cloud agnosticism" is not something that the company wants to invest in... and when developers who try to abstract those interfaces in order to loosely couple to the vendor were actively penalized while developers who rush to ship spaghetti code are rewarded, then the cost of switching off of cloud will be close to the cost of an entire ground-up rewrite of the code-base plus the added expense of having an IT staff bootstrap the bare metal infrastructure while coming up with replacements for things that the cloud just gives you out of the box (like auto-scaling and automatic redundancy promotion in failures etc.).
I worked for a startup a few years ago where the CTO was heavily invested in AWS. We were at the very early stages and could have switched fairly easily had we wanted to. But when I made certain recommendations to help mitigate the inevitable vendor-lock-in the CTO's response was "why on earth would we EVER switch off of AWS? This is a non-issue." If I brought up rising costs over time the answer was "we can always negotiate lower costs if we have to." It just seemed so far fetched that the company would ever even consider an alternative that all recommendations to take measures to mitigate against vendor lock-in were dismissed as unreasonable concerns.
New start-ups will begin to examine whether cloud is a given or whether or not bare metal makes sense. But existing businesses? The cost of switching is way too high and that's by design.
Re: (Score:2)
Re: (Score:2)
Because companies are making leaps and bounds in advances in locking people in. Most of the significant technological advances recently have been made in screwing over the customer.
Re: (Score:2)
Because on premise has its own cost, and it's not just money. You need to hire admins, you need to build out physical infrastructure including racks, cooling, and network hookups. You need physical rented space. You need to accurately predict growth to buy hardware ahead of needing it. The benefit of the cloud is that it's easy, and a lot of mistakes you can make are easily fixable by spending more. Even if the cloud is slightly more expensive its still frequently the right choice. Your scale needs to
Re: and they vendor lock you in to make it hard to (Score:3)
With the right tools you can hop between AWS, GCP, and Azure. And yoy used to be able to drop VMsphere into your own datacenters and run K8s and full VMs together. Although the new owner is likely going to screw up that business.
Re: (Score:2)
Re: and they vendor lock you in to make it hard t (Score:2)
Generally less capital required to do it the expensive way. We rent laser printers and logic analyzers too.
Re: and they vendor lock you in to make it hard (Score:2)
We rented a fancy network testing tool once but that was the one time in 10 years we changed the cabling due to a building renovation
But Logic analyzers? Seriously?
Re: (Score:2)
Yea. we have a LOT of logic analyzers. Some with some rather pricey modules and associated software licenses installed. It's an accounting thing. If you spend money on a reoccurring cost it's different than acquiring an asset. Plus we don't need to depreciate our LAs, just return them at the end of the lease and get a newer one.
Re: and they vendor lock you in to make it hard t (Score:2)
Containers do somewhat improve security by isolating things and can auto-loadbalance but other than that I agree
Whatsapp used to run on 2 BSD boxes with Erlang
Re: (Score:2)
Yeah it was obvious, but managers in their arrogance know so much better than the people who actually do the work. But at least the look like they are doing something productive.
Re: (Score:2)
Really, I don't think it was a good idea either (Score:2, Interesting)
Re: (Score:2)
I see I'm not alone in thinking The Cloud was a scam to begin with, just too many haters hanging on Slashdot today.
Re: (Score:2)
Too many haters on /. eh? Guess you haven't been around /. for a while.
The first rule of /. is: Being a hater (of [pick any topic]) is de rigueur.
The second rule of /. is: If you plan on being a visible hater, but you don't want to damage your social credit score , then learn to post Anonymously.
The third rule of /. is: Learn to properly format your posts AND turn off dat dam SmartQuote feature on your phone.
Most /. posters never seem to get to rule #3. Guess they can't read that far.
No. (Score:5, Insightful)
Admins never thought it was. They could (and did) give you a laundry list of potential problems that would come with adopting the idea.
The idea was popular with finance and executive types looking to reduce capital expenditures and to reduce headcount.
Re: (Score:2)
I'll add that I think it depends on the use. A "no" (often?) means it was an inappropriate use, probably picked for the wrong reasons, by someone other than the admins, probably higher up on the food chain, who may not really understand how things actually work.
Re: (Score:3)
Every admin thinks they're good enough when it comes to security. The truth is admins are a mixed bag, and few live up to how good they think they are. Compare that to AWS, which only has one very large instance to manage and Amazon resources to make it the best. With cloud you get some of the best security people in the business, and all you have to do as an organ
Re: (Score:2)
It doesn't matter how good the lock is if the key is under the mat.
Re: (Score:3)
So to use your analogy, it does matter how good the lock is if you are good about managing your keys.
Re: (Score:3)
Keeping things in house doesn't have probably billions of dollars of marketing to try to fool managers into thinking it a good idea.
You maybe right with liability reduction you have someone to point the finger at if things go wrong, its not they way it should be though. I don't think you are right with security though you really don't know what the security is like at these companies. Its like saying the security I don't know is better than the security I do. All you have is plausible deniability.
Re:No. (Score:5, Interesting)
This, a million times this. At one of my employers there was a question of buying a few servers to replace old junk, or moving to cloud. After I did some math and found that compared to the pricing of the cloud, our own hardware would pay for itself in a year. Begrudgingly, the leadership went my way - they could not let go of the idea of the cloud being a magic solution to everything, but they could not find fault in my reasoning either.
For some other services that did go to the cloud, leadership has been bewildered multiple times that the cloud ended up with a higher TCO than self-hosting. I on the other hand am bewildered that the leadership is unable to understand the very basic principle of "there ain't no such thing as a free lunch".
There seems to be at least some sense to this, though. One of my bosses explained that sometimes, the question is ROI. Drop that upfront money into hardware, or hire another developer for a year. With the income from the developer, you could pay extra for the cloud and still come out ahead. So there a bit more to this than the surface level considerations, if you weigh your options. But looking at the general state of the industry, with everyone approaching the issue from the TCO pipe dreams side, it seems that the most bosses are unable to grok even the surface level, much less so figure out what's the best move to be found...
Cloud services are wasteful uses of resources (Score:2)
Re: (Score:2)
Re: (Score:3)
Admins never thought it was.
To add to this Admins are the anti-thesis of the cloud. The whole benefit of someone else's computer is that they are likely better at running it and maintaining it than you are. If you have qualified admins on staff then one of the major benefits of the cloud is defeated - you have your own capability and don't need to rely on someone else.
Re: (Score:2)
I dunno, I made a lot of money moving things to "the cloud". It looks deceptively simple, and the vendors really play on that aspect of it, but if you want it to be secure and manageable, it's a lot less simple than it seems.
I've been called into a few shops where they put down the company credit card, someone spun up a couple of servers by waving the mouse and then found it wasn't "magically" that easy to do whatever else they wanted to do. I came along and laid things out in some sort of sensible manner,
The cloud breaks Moore's Law (Score:2)
Costs are supposed to go DOWN as technology increases.
Re:The cloud breaks Moore's Law (Score:5, Insightful)
Applying Moore's Law to this (cloud costs going up faster than inflation) is like applying Occam's Razor to shaving.
Re:The cloud breaks Moore's Law (Score:4, Funny)
Personally, I use Occam's Razor to cut costs.
Re: (Score:2)
The problem is that when people move to the cloud, it means fewer devices (servers, SAN appliances, and so on), which breaks economy of scale. This kills Moore's law because each device now has to be more expensive.
The cloud and centralization will kill all progress because of this. Having everything decentralized is why we can have decent CPUs, as opposed to companies charging megabucks for mainframe-tier CPUs that are only used by a relative few.
Re: (Score:2)
If you claim most modern CPUs are not “decent” I call bullshit. Your cellphone has a better CPU than a supercomputer did 15 years ago. The server CPUs that are used by relatively few have a totally different design and optimization envelope than anything you might want to have for personal use.
Re: (Score:2)
You might notice that the MBAs are in charge now, at least at the big cloud providers (Amazon, Google, Microsoft). For an MBA, if you don't have constant, double-digit growth, you're a failed business. Numbers must go up, every quarter. The only customer they're obsessed with is the shareholder.
Cloud costs are just going to keep increasing. You're locked in, what are you going to do, go to a competitor who's also co-incidentally (wink wink) raising their prices constantly as well?
See also streaming services
Re: (Score:2)
If I could stand constantly selling myself, I'd run a consultancy for setting up "on-premises cloud" (ie, servers) for small and medium businesses. Ah well.
Here here! Maybe a bunch of like minded people can/will setup a non-profit to provide those services? Like, "The Anti-Cloud Group," or something. That would get around the shareholder demands for ever increasing profits, which any publicly traded company is more-or-less required to do. It would probably end up having its own server farm (cloud) as well, but manage it like we all did for ages - real cost based pricing and managing things to optimize for that, as well as straight forward options for dedicated
Re: (Score:2)
I'm thinking of doing it myself. Starting with my own data driven resume, which is currently hosted in Azure.
My first step is switching my internet from cable to a fiber provider, at half the cost and twice the bandwidth.
The Squeeze (Score:5, Informative)
Almost like competition is important... (Score:2)
Business 101 says to create a monopoly. Look into monopoly protections... at least here in the US they're much more vague/minimal than I'd assumed (though based on life experience I should have known).
Oh, and that justice can't happen when you need to be rich to get it. And that public perception matters more than it should... which can be bought and/or manipulated often.
Did they account for inflation? (Score:2)
Just looking things up, apparently 2022 to 2024 is 7.5% inflation rate annually, and the Azure cloud was 14%.
So I'd argue the costs mostly remained level, and I have not been that impressed with computer technology improvements lately. IE costs aren't dropping as much as I'd like.
Re: (Score:2)
still higher then the 3% raise you would have paid the local admins and the hardware cost that was $$'s 3 years ago.
no one ever got fired for buying cloud... maybe they should.
Depends on if you were paying admins (Score:3)
One of the promises of the cloud is that whomever is maintaining "the cloud" is also installing, racking, and managing firmware on server and network hardware as well as patching operating systems for you, so that you don't have to pay staff to do those tasks (and therefore it is cheaper because you can share that cost with other cloud customers.) In addition, the theory is that the people managing the hardware, network, and some level of redundancy and disaster recovery planning at the cloud provider were able to do it with more skill and reliability, and with lower cost than you could/were doing on your own.
It's a nice concept in theory and probably mostly so in practice, but if you weren't doing all that originally (or you weren't paying your people to do that as part of their official duties), then you are going to end up paying for something that you were not before - it costs more because you are getting services that you were not paying for before.
If you *were* doing that stuff before internally, you were probably not doing it as a profit center as well. Cloud providers as a business are going to try to make a profit.
Re: (Score:2)
There's kind of a mix - we have a systems reliability team that are kept pretty busy. Cloud providers don't roll all your thousands of instances when they have a new OS image to deploy. There's still a lot of sys admin stuff to be done. But we don't have a hardware team traveling the world installing blades in data center cages anymore.
Think (Score:2)
The honest answer is that choosing cloud vs not cloud is highly dependent on workload and individual company circumstances. Cloud won't always be cheaper, on prem won't always be cheaper
Re: (Score:2)
+ how many of these migrations were lift-shift and didn't take advantage of cloud flexibility. If you migrated and still treat it like on-prem, yes it'll definitely cost more. Cloud isn't the same as on-prem. If your 'migration' was essentially just "rsync *" then you're going to have a bad time.
Re: (Score:2)
Cloud has increased at double the inflation rate. The servers you would have bought last year are still running this year, no need to re-buy. Payroll is probably up, but few payrolls even keep up with inflation, much less double it. Power bills have gone up a bit depending on the area, but probably not 14%.
Next year, the servers will still be running just fine.
Re: (Score:2)
> Hardware prices have gone down, while performance increased, right? And electricity costs? Cooling costs? But surely labor costs for highly talented IT personnel have gone down right?
You're still paying for all that PLUS corporate profit overhead, and possibly increased network bandwidth as well. All for the privilege of putting all your business's eggs in someone else's basket.
=Smidge=
Re: Think (Score:2)
Re: (Score:2)
and now that you moved all your crap to cloud and lost the talent that maintained on prem, getting back is not even possible.
once you go cloud, you can never go back (mostly)
Anyone who thought cloud was good (Score:2)
Anyone who thought cloud (as it is used in this context) was good was an idiot. It was a money grab and a scam from the start. Why would anyone rent their software and have it stop working if you decide not to pay anymore. Now SUPPORT is truly a service and something that no one would argue that you pay for it on an ongoing basis, but not software, and definitely not hardware. SAAS or HAAS is and always has been evil. Cloud is not bad as long as it is an in house cloud service and not rented.
"Admins wonder if the cloud was such a good idea" (Score:3)
Uh, I was a System Administrator for over a decade before they started calling me an Engineer. What admins are they talking to, the ones that have lost their jobs because cloud resources are considerably simpler to manage and maintain than naked iron, or the ones who have discovered they get paid less than they used to because even recovering from disasters is way simpler if someone else manages the hardware layer.
The hardware layer. *Hunh* Sounds like I'm talking about a network. Next, we'll talk about abstracted hardware, abstracted storage, abstracted network configurations, the OS layer, the software layer, the interface layer . . .
This pendulum will keep swinging around forever (Score:3)
Dumb terminal-to-mainframe, personal computers, server-client ("smart" clients), cloud computing & (originally underpowered) phones, computing on phones (and in browsers generally with WASM), computer history doesn't repeat exactly but it sure does rhyme a lot. (Also see synchronous/asynchronous, and other related pendulums.)
Of course for any given challenge the requirements, and the tools available, vary both in general over time, and depending on many variables in the specific context.
Re: (Score:2)
If it makes them money to charge stuff for reinventing something existing, then I'm guessing it becomes the latest cool thing. We started with terminals, went to computers, then to X-stations, back to graphical workstations, then to JavaStations, back to PCs, then to the cloud, and now with AI stuff, stuff is heading away from the cloud because the advertisers want their data processed on people's PCs rather than have to do it with their own CPU cycles, and glorious stuff like MS Recall (which will be a bo
There is no surprise here (Score:2)
It was clear that a) the cloud would be expensive and b) the cloud would not live up to its promises of reliability and security. As usual, only the clueless are surprised by this.
Always remember.... (Score:4, Insightful)
Re:Always remember.... (Score:4, Insightful)
It's a bit more than this, the main advantage of the cloud are the "always on" services and "processing power on demand", something that is not (easily) possible in a smaller/dedicated hosting, unless you want to invest in the same amount of resources (not very likely for most customers).
Re: (Score:2)
and you likely care a lot more about your data than they do.
If you're a Fortune 500 company then you may be right, but in a huge number of cases you're dead wrong. The move to the cloud is brought to you by bean-counters, the same bean counters who see IT admin as a cost centre to be minimised. Many MANY companies out there are woefully under skilled at maintaining their hardware and data security. We talk about these companies constantly here.
Your point doesn't go against the concept of the cloud. If you care about your data one way of looking after it is to ensure
Conditionally Stupid (Score:2)
If you don't need the Internet to operate (if your data is all in house already), and you can afford proper IT infrastructure, then giving up your systems to someone else means you're giving up control for no reason.
Infrastructure is cheap (Score:2)
Infrastructure is cheap. I personally oversaw one client with wasteful infrastructure (VERY overprovisioned machines), but it was only 50k/yr infrastructure spend against $1mil+ revenue.
Managing the infrastructure yourself will be a lot more expensive, especially if you plan on resilient architecture that AWS offers. Sure, vendor lock is ugly and can get you in a potentially difficult situation in the future with rising prices, but with the competition in the cloud computing landscape, I wouldn't worry too
Re: (Score:2)
this is the answer. It seems like many a slashdotter here thinks everything is just "hosting a website".
There are a lot of services you need to manage even to publish just a static server. And you have to keep up with security patches.
DDoS on your internet connection? tough luck. Sudden spike in traffic? Tough luck (this IS slashdot, we all remember "the slashdot effect"). need a CDN? better learn how to set one up. and so on.
When your revenue is in the several million and you can keep a dedicated team for
Re: (Score:2)
If there's anything an organization should be worrying about with cloud infrastructure, it's security--who has access to your data--not cost.
Given the almost weekly Azure outages now, accessibility has got to rate right up there as well as something to worry about.
Out with the old (Score:2)
I wonder... (Score:2)
if the hype over AI has the same timeline.
Lift and shift; Adapt; Save (Score:3)
Re: (Score:3)
but their head count is generally smaller
This was happening already. The same people who brought you the idea of saving money in the cloud were the ones who saw IT administration as a cost centre to be minimised. After years of gutting IT it's no wonder then that we hear a never ending string of stories of security and data failures in businesses.
The thing is the cloud is someone else's computer, and that someone's core competence is looking after it, while your own CFO is specialising in letting your own infrastructure rot.
Re: (Score:2)
My clients IT too, two months to answer an email. The other client has a five day process to add a DNS entry. Frankly, their head counts deserve to be reduced.
Dynamics 365 (Score:2)
Microsoft has for years been pushing its on-prem ERP customers to the new cloud solution, Dynamics 365. The promise is occasional feature updates and frequent quality (ie. bug fix) updates. All well and good; a modern ERP system is the lifeblood of the company.
Microsoft recently announced that they were "retiring" (ie, decided not to pay the license costs for) payroll tax in the US. Now, this may sound like a major feature update. Microsoft, in their wisdom, thought otherwise. In a quality update over the r
Are they doing it right? (Score:2)
However most organizations i see are doing very little more than data warehousing, and not really utilizing the cloud for what it should be good for such as centralized processing of functions while distributing information to a large number of users. If you're just storing data, then yes the cloud is kind of overkill. I wonder how much of this is just the wrong tool for the job.
The Cycle of New Stuff (Score:2)
We've seen this before.
Netflix, Uber, Airbnb, cloud providers, Adobe, ... everyone follows the playbook.
You must be kidding (Score:3)
No (Score:2)
Web based apps are fine (Score:2)
I remember going from desktop to web based apps and how much less support there was. Computers and their libraries break *constantly*. Just not having to spend hours dealing with broken
Re: Web based apps are fine (Score:2)
So use a native compiled language with statically compiled libraries. Problem solved. Each runtime dependency you add just increases the amount of shit you have to deal with that much more.
Stupid question time (Score:2)
It does seem like lazy journalism to omit the time period from the reporting, but I guess at least they didn't bother tell us how many jumbo jets 200GB will fill. Sorry, that was a ton
Private cloud is a good idea (Score:2)
Commerc
No waaaay! (Score:2)
Don't put anything into the cloud that you can't get back out in two hours.
I have yet to meet a professional admin who didn't follow that rule.
This is considered basic common sense among professionals.
This is "the way" (Score:2)
If that's not "part of the plan", then you'll be surprised by your bill.
Good... hope they all get gouged.... (Score:4, Interesting)
It's really been a bad thing for I.T. careers on the whole, shoveling everything into the cloud. It's caused many smaller to mid-sized businesses to eliminate their sysadmin/network admin staff or at least to cut things back so one guy is tasked with what's left of that, tacked on to a PC desk-side support role (with no increased pay).
I mean, at some point, cloudifying things means the only job roles you're still hiring for are for your client PC laptops and workstations that connect to it.
Just follow some of the Reddit forums like "sysadmin" or "it" for a little while and see all the stories posted there. It starts painting a picture of how things have changed - and it's not really good. Just like things consolidated so everyone's using AWS, Azure or Google for cloud application/server hosting? The jobs maintaining the servers are consolidating too.
It's always been cautioned that the cloud is nothing more than paying someone else to handle your data on their computers for you. But those parties are free to charge you as much as they like for the service, AND you really have no recourse when things go down. (Not like Microsoft cares about your specific business and how their outage affects it. You just get to sit and twiddle your thumbs like everyone else until they sort it out.) The bean counters advocated this stuff mostly because jettisoning employees as part of it made their numbers look good. But there are downsides they ignored or downplayed.
Over half reported a rise in their cloud bill (Score:2)
Well duh, if companies are moving from on-prem to the cloud, of *coursed* their cloud bills will increase.
Cost has to be managed differently (Score:2)
A lot of shops that went from on-prem to cloud, didn't really realize how to manage costs in the cloud. They were used to buying "whole" computers. Because they couldn't easily be "scaled up" as needed, they bought more than they needed. That kind of thinking often carried over into the cloud. Companies I've worked for routinely stood up pretty beefy cloud VMs, when they only needed a couple of vCPUs.
At one company, our monthly Azure bill was $20K per month. When we started to look at where the money was go
Re: (Score:2)
A lot of shops that went from on-prem to cloud, didn't really realize how to manage costs in the cloud.
Precisely this. Simply replicating everything with no consideration to actual workload or function is the primary reason for the billing surprise.
Taking into account hardware replacement, management, power, facilities costs and patching our bill remained relatively static despite expanding our internal service offerings following our migration - Improved DR and auditing capability along with significantly improved security posture.
This is ignoring all the automation, flexibility, adaptability and deployabil
Most on-prem admins cut costs where they shouldn't (Score:2)
For example, many on-prem admins don't do full disaster recovery planning and preparedness. Many can't even prove that their database backups actually worked. They performed the backup, and just assume it was good. And the backup is stored in the same room as the original.
In the cloud, it's a lot easier to get executives to pay that monthly rent for that geo-redundant server, than to get them to pay big bucks for provisioning a second or third redundant "on prem" data center.
After? Try "Before" (Score:2)
After an initial euphoric rush to the cloud, admins are questioning the value and promise of the tech giant's services.
Every admin I knew had questioned the value of "The Cloud" the moment they heard about it. Competent admins cannot see the value of putting the company's services in another company's machines over running in machines inside the company.
The C-suite, however, get the euphoric rush for the $$$$ dollars signs they can see in their bonus from "cutting costs" by moving to the cloud, whatever comes after a few years later is not their problem since they are already gone.
Yes cloud is a great idea.. (Score:2)
.. for our company.
(We sell cloud services at a high profit)
Who would have thunk? (Score:2)
I just LMAO. This is a surprise?
I use clouds all the time (yes, plural, not entirely by choice). Why? It is (well, can be) very convenient, it provides short-term security, it allows me to bind in people less technically inclined, it makes shifting data between platforms easy. In the end convenience. Which is bad counsel for IT stuff ....
I also assume that any (or all) of them can "fail" at any time. Read "fail" as: actually fail, loose part of or all my data, corrupt part of or all my data (oh, that has ha
Pros and cons (Score:2)
Pros and cons of cloud:
Pros:
your salesman is happy
you worry about backups
Cons:
your salesman is happy
you don't worry about backups
your data is literally in the hands of ANYONE else and you've no way to know
if anything goes wrong, you're fuct
Pros and cons of local servers:
Pros:
your data is there even when the internet isn't.
speed/connectivity is at its highest
No cloud fees
Cons:
you might need to hire an IT guy to manage it
Re: (Score:2)
Exactly. On-prem and cloud aren't mutually exclusive. Sometimes there are benefits both ways.
For example, we provide services to companies - we have software that can run on-prem, but lots of companies we deal with just do not have the IT resources to deal with needing to install management servers in-house, firewalls, and data protection that comes with it. They may not even have infrastructure with the necessary uptime they need. So the cloud hosting makes complete sense to them - they do not have full ti