Proof that software as service/cloud based, will never work for long term ...

[tom66]

I 'd guess that many don't know that we already had cloud services / SaaS in the early 2000s. But it wasn't successful back then. A typical setup was based on Citrix plus standard applications the customer could rent. The customer used inexpensive thin clients or normal PCs (fat client) to access the platform. Companies offering that service were called ASPs (Application Service Provider).

AWS was started in 2002, using the term web services, and as the compute people noticed the telecoms people using the term "cloud" a lot, AWS started calling their offering cloud computing in 2006. So, I think we all know there were cloud services in the early 2000s. Cloud like services have been widespread since the shared data centre services of the late 1990s. Its mostly the names which change.

Nothing fundamentally wrong with cloud computing.  IT people who do on-premises work are often skeptical because it puts them potentially out of a job.  But then they can just work on maintaining cloud stuff, it's just that the responsibility for the hardware and infrastructure is someone else's, so they may feel their skills are less useful if they have decades replacing worn disk drives and debugging borked kernels, and now it all runs on Docker or something like that, so it's really no hassle to spin it all back up again on a freshly imaged server.

--

The problem with consumer cloud computing is when it integrates heavily with software-as-a-service and there's no guarantee of long term support.  If you're a big enterprise you can afford  to sign an agreement with AWS/Azure/etc and an IT systems consultancy for some eye-watering sum of money and there's an explicit guarantee of services being available for, say, 10 years, and migration processes when things change.  But as an ordinary consumer you have none of that.  You're not guaranteed that the hardware you buy will be supported long term on the cloud platform.  Your $10 a month, or whatever, is not necessarily enough to keep the company interested. 

I think the only way this can be improved is by legislation and universalisation.  The cloud products need to be able to connect to anyone's cloud (ideally even a self-hosted one) and that cloud service becomes a product you procure and competition guarantees price/service/availability just like the cellular network market is (usually) competitive.  If Google's consumer cloud for Nest products disappears, no big deal.  You can go somewhere else.  The self-hosted cloud is the most practical one IMO, because a simple piece of home hardware and a way to route that to the internet gives you full automation.   (This is why I use Home Assistant for my home automation tasks).  Matter and Zigbee also promise to universalise some home automation stuff, but there are still proprietary elements that are painful.  My advice would be not to purchase any cloud connected home automation kit unless it has some way to run as well as required from a non-cloud system.  If the cloud functions add some convenience but were not a major reason for purchase maybe that's ok, but I'm still not a fan.

[Nominal Animal]

IT people who do on-premises work are often skeptical because it puts them potentially out of a job.

No, that's a lie.  I was one for a few years.  Nobody worries about losing their job when moving to cloud; only about their job becoming impossible to do well.

The same sysadmin work has to be done regardless of where the hardware is; nobody is out of a job when shifting from local to 'cloud'.  Everyone I know considers hardware maintenance a burden they'd happily offload to someone else, especially because it is so darned hard to get enough funding to have spares on hand.

The skeptical ones I know (doing HPC and computing clusters) are so because they become responsible for stuff they cannot control, and worry about the connectivity and bandwidth to the cloud.  (Not many worry about data leaks/thefts, because they do research; I do/did.)

Those who maintain web services and databases don't mind the cloud, as virtual hosting with uptime guarantees makes their work easier, especially when larger sites can be split into subdomains and scaled as needed.  (The only exception is contracts: many fear they are done at the CXX level without proper domain knowledge or understanding the requirements and complexities at all.  With a bad contract, it can be an utter nightmare.)

[tom66]

IT people who do on-premises work are often skeptical because it puts them potentially out of a job.

No, that's a lie.  I was one for a few years.  Nobody worries about losing their job when moving to cloud; only about their job becoming impossible to do well.

The same sysadmin work has to be done regardless of where the hardware is; nobody is out of a job when shifting from local to 'cloud'.  Everyone I know considers hardware maintenance a burden they'd happily offload to someone else, especially because it is so darned hard to get enough funding to have spares on hand.

The skeptical ones I know (doing HPC and computing clusters) are so because they become responsible for stuff they cannot control, and worry about the connectivity and bandwidth to the cloud.  (Not many worry about data leaks/thefts, because they do research; I do/did.)

Those who maintain web services and databases don't mind the cloud, as virtual hosting with uptime guarantees makes their work easier, especially when larger sites can be split into subdomains and scaled as needed.  (The only exception is contracts: many fear they are done at the CXX level without proper domain knowledge or understanding the requirements and complexities at all.  With a bad contract, it can be an utter nightmare.)

Well, we must know different people .  I've met a guy who is extremely skeptical of cloud computing but when pressed it ultimately came down to his being made redundant and not being able to cope with everything being in a cloud portal with different management tools and certifications required.  Being an older gentleman he just took early retirement rather than re-skill, and I am sure this is not uncommon. 

If you read "The Register", a UK journal of IT stuff, you'll find that a great number of commentators on there are extremely skeptical over cloud computing particularly due to losing access to physical hardware.  Yes there are also other entirely legitimate reasons to be skeptical but there seem to be just as many obsolescence concerns.

[Nominal Animal]

Well, we must know different people .

I did know one "webmaster" who in 1999 did not know what a "server certificate" was, but the rest I've worked with loved making things possible for others (at universities and research organizations in particular).

a great number of commentators on there are extremely skeptical over cloud computing particularly due to losing access to physical hardware

I understand the problem of losing control over the underlying hardware while still being responsible for the services to be available; but, among the dozen or two I know sufficiently well to say, not one minds handing over the hardware maintenance to someone else.  Those who object to cloud move object only in specific cases, too; not categorically.

I guess there could be people out there doing only Windows IT support and replacing hardware that are afraid of having to retrain when moving to cloud; I don't know of any, though.  I only object to the idea of all/most IT people doing on-premises work fearing the cloud move puts them out of a job, the way you put it.

[paulca]

As someone in that position right now, working in a company who pride themselves on their "Cloud Transformations", there are many things which people assume will cost jobs.  This usually appears to end up wrong.  However.  What does happen, which is not accounted for in those studies is "career displacement and turbulence".  This is particularly impacting on the more senior and more experienced engineers.

It's a bit like watching climate change happen out to 2100, but as everyone adapts and migrates around away from hotter parts to the now unfrozen poles the population doesn't decrease dramatically so it's okay then?  Yes?  No.  It misses the turmoil of displacement, migration and many will fall along the way.

Cloud means many, many things.  The "Cloud" I am talking about here is basically reselling AWS and Azure products.

AWS and Azure don't just "rent out" virtualised hardware, the provide massive swaths of "Codeless", "Serverless", "Application in a box", "Network in a box", "Security in a box" type components.

Developing one of these projects has nothing to do with software engineering.  Nothing.  It's just "Using a web application provided by AWS/Azure to configure a set of business components and applications also provided by AWS or Azure.

When you find something lower level to peak your interest you, it is usually found in integrating legacy systems into the new cloud arch, or migrating them. 

The trouble is.  It creates lots of jobs for the inexperienced "engineers" they pluck off the street with nothing buy an Azure certification.

Being an AWS Developer or an Azure developer and that is the headline on your CV...  you are not an engineer.  You are an application operator.  You are not much more than an office clerk using Microsoft Office.

The work is boring, dull, and worthless.  ALL of the components you will be using will be actual "real" open source components, but ALL of them will be abstracted away and renamed, protected behind APIs which force you to do things in the "AWS Way" or the "Azure way".  Of course for every Az/AWS service you consume you pay for it.  However, to intergrate outside the cloud costs twice as much.  These cloud guys are smart.  They know their market.

The trouble is that market does not know it's own business.  When they do finally atriphy away all core engineers from their upper ranks and are left with 99% business goons in the upper half and the lower half is 99% AWS Home trained button pushers...  they are 100% dependant on AWS.  AWS (et. al) know this.  It's exactly what they intend.

So if you actually want to develop software and not just "configure and integration" other peoples canned solutions, the number of sectors and businesses you can work in is diminishing rapidly.

You could go work for AWS/Azure, but, that for some reason makes me feel ill.

So right now I have some serious questions on the sustainability of my career right now.  I don't know if my skills will not become neiche in the coming years. 

Some people argue this is just natural progression.  It's just the tech landscape mutating as it develops.  You just need to keep learning the new things right?  What has changed?

Cloud.  Cloud is different.  As I said above, they are not platforms designed to help you engineer software, they are platforms designed so that you don't need to engineer software.

While "Software Engineering" is a self-deprecating endeavour - we write software such that non engineers can do things the previous required engineers to do - I think this rapid "transformation to cloud" is going to move the market for actual "Software Engineers" out of the majority of "Business solution provider" type companies and SIs.  Actually software engineering will move into more "neiche" markets and sectors and in provision of those cloud components within MS/IBM/AWS etc.

I have 11 years ideally.  More realistically 18 years before I can retire.  At the moment, I do not hold high hopes

[tom66]

A close family member works with AWS a lot (he's a senior management consultant).  There is no shortage of money in his field.  Most of the engineers that work under him have a salary in excess of £100k (usually the job is fully remote as well).  It is not quite as simple as you put, "low code" does not mean "low engineering".  There is still a great deal of brainpower required to build and run these systems and when they win a contract of, say, £50m to supply a major business, about half of that is paying for the engineering fees to design and keep the system going.  And this is a super competitive market with typically 3-5 other consultancies bidding against a tender. 

It is true that it encourages strong vendor lock-in though, so whether these companies are making the right long term decision is another matter.  There's a strong thread of IT incompetence in large organisations which has perhaps promoted the outsourcing of this technology, rather than investing internally, because that's difficult and requires a longer term vision.  I'm thinking of the British Airways annual system crashes which result in cancelled flights - they're now desperately migrating everything to AWS so I guess we'll see if this gets better.

[Nominal Animal]

There's a strong thread of IT incompetence in large organisations which has perhaps promoted the outsourcing of this technology, rather than investing internally, because that's difficult and requires a longer term vision.

Agreed.  Plus, outsourcing also provides a perfect scapegoat if the vision turns out to be crap.  That is, it is 'safer' for an executive to outsource, than to invest the same amount internally, because it is much easier to deflect blame to the vendor, than to the underlings one is in charge of.

[paulca]

The AWS, Azure, Blue Ocean et. al. lock in is very real.

I found out to my confusion recently that the UK government is single vendor locked now.  To Azure.

As far as my historical understanding of government IT ... it should NEVER be single vendor.

[tom66]

Well, the biggest problem with government contracting is that it is even contracted out altogether.

These huge projects should be done in-house with a team of engineers employed by the agency.  The agency should be paying market rates to attract the best talent and the management is selected for engineering expertise rather than the ability to suck up to higher ups.

But we all know that will never happen.

[Nominal Animal]

These huge projects should be done in-house with a team of engineers employed by the agency.

In Finland, huge IT projects have about 33% chance of yielding an usable product, because of leadership and contract issues.  Essentially, outsourcing companies are gleefully exploiting the poor contracting and specification skills, for maximum profit.  It is so bad that even if you run a project that wastes millions, your career won't suffer, because "it happens to everyone".  Often management still gets their bonuses even if the project fails, "because they did their part so well".

It is not a grassroots problem, it is a management, executive level problem.  I know a lot of grassroots people who could do the technical stuff, easy, but the problem is getting qualified management to run such projects.

To me, it looks like outsourcing to cloud and software as a service are mostly a way for poor management to externalize their own responsibilities of the result.  That is, it is a perfect opportunity to get all the glory if something works, while perfectly deflecting any blame when it doesn't.  After all, it is always somebody elses money.  Puffing up the middle management part, distancing the decision makers as far as possible from the people who are supposed to do the actual work, is very important for that.  It also explains why organizations tend to grow in the middle-management level the most.

[nctnico]

There's a strong thread of IT incompetence in large organisations which has perhaps promoted the outsourcing of this technology, rather than investing internally, because that's difficult and requires a longer term vision.

Agreed.  Plus, outsourcing also provides a perfect scapegoat if the vision turns out to be crap.  That is, it is 'safer' for an executive to outsource, than to invest the same amount internally, because it is much easier to deflect blame to the vendor, than to the underlings one is in charge of.

That is not the reason. The real reason is cashflow and balance. Any money a company spends on having servers in-house is money sitting on the floor as dead weight depreciating like crazy as well. A typical company is much better off spending that money on inventory and/or people who can push out a product to earn an ROI.

[tom66]

There's a strong thread of IT incompetence in large organisations which has perhaps promoted the outsourcing of this technology, rather than investing internally, because that's difficult and requires a longer term vision.

Agreed.  Plus, outsourcing also provides a perfect scapegoat if the vision turns out to be crap.  That is, it is 'safer' for an executive to outsource, than to invest the same amount internally, because it is much easier to deflect blame to the vendor, than to the underlings one is in charge of.

That is not the reason. The real reason is cashflow and balance. Any money a company spends on having servers in-house is money sitting on the floor as dead weight depreciating like crazy as well. A typical company is much better off spending that money on inventory and/or people who can push out a product to earn an ROI.

That's very much dependent on the scale of such resources.

I would argue a company like an airline actually will make good use of their server resource and so the asset investment is worthwhile. 

The problem as ever is the lack of skills or the ability to find the right people to run IT at your organisation, hence outsourcing to companies that specialise, but at great expense.

[nctnico]

IMHO succesfull outsourcing starts with having people on your side (either employed or contracted) who have the knowledge define the requirements, help selecting suppliers and to judge whether a supplier is doing a good job or not once the project commences. The latter takes effort at both ends BTW especially when rolling out a new IT system.

[coppice]

IMHO succesfull outsourcing starts with having people on your side (either employed or contracted) who have the knowledge define the requirements, help selecting suppliers and to judge whether a supplier is doing a good job or not once the project commences. The latter takes effort at both ends BTW especially when rolling out a new IT system.

Successful outsourcing starts with carefully assessing the incentives of all involved. Tell me the incentives and I have a high probability of predicting the outcome.

[Nominal Animal]

There's a strong thread of IT incompetence in large organisations which has perhaps promoted the outsourcing of this technology, rather than investing internally, because that's difficult and requires a longer term vision.

Agreed.  Plus, outsourcing also provides a perfect scapegoat if the vision turns out to be crap.  That is, it is 'safer' for an executive to outsource, than to invest the same amount internally, because it is much easier to deflect blame to the vendor, than to the underlings one is in charge of.

That is not the reason. The real reason is cashflow and balance. Any money a company spends on having servers in-house is money sitting on the floor as dead weight depreciating like crazy as well. A typical company is much better off spending that money on inventory and/or people who can push out a product to earn an ROI.

No, you confuse hardware with outsourcing services and entire systems.  You don't have to have the servers in-house.  Like I said earlier, the IT devs/admins/people I know would be happy to outsource the hardware.  It is outsourcing the control that is the problem.

From a business perspective, the solution that in the long term produces the best return of investment, makes the most sense.

However, in current large organizations typical managers don't stay in the same position more than a couple of years, so anything beyond that is irrelevant to them.  What matters to them, is quick returns and being promoted before the shit hits the fan.

[SiliconWizard]

There's a strong thread of IT incompetence in large organisations which has perhaps promoted the outsourcing of this technology, rather than investing internally, because that's difficult and requires a longer term vision.

Agreed.  Plus, outsourcing also provides a perfect scapegoat if the vision turns out to be crap.  That is, it is 'safer' for an executive to outsource, than to invest the same amount internally, because it is much easier to deflect blame to the vendor, than to the underlings one is in charge of.

Yes, and beyond costs, this is the #1 reason organizations choose to outsource anything in the first place.

[madires]

Something new: Oopsie, we've just deleted your account!

Google Cloud accidentally deletes UniSuper’s online account due to ‘unprecedented misconfiguration’: https://www.theguardian.com/australia-news/article/2024/may/09/unisuper-google-cloud-issue-account-access

The online account included all cloud services and backup. UniSuper is just a small company managing about US$ 125bn for 620k customers. Luckily they have another backup somewhere else.

[SiliconWizard]

Something new: Oopsie, we've just deleted your account!

Google Cloud accidentally deletes UniSuper’s online account due to ‘unprecedented misconfiguration’: https://www.theguardian.com/australia-news/article/2024/may/09/unisuper-google-cloud-issue-account-access

The online account included all cloud services and backup. UniSuper is just a small company managing about US$ 125bn for 620k customers. Luckily they have another backup somewhere else.

“This is an isolated, ‘one-of-a-kind occurrence’

Yeah. Sure.

[Nominal Animal]

Isn't every single incident an 'isolated', 'one-of-a-kind occurrence'?

If it is a series of events, it's still an incident, just spread over longer period of time.  Thus, every incident is unique and isolated from others.

If different variants occur more than once, then it is no longer an incident but a feature of the service, by business conventions.

[MrMobodies]

Just came across this article:
https://www.nytimes.com/wirecutter/blog/security-camera-cloud-storage/

We suggest that most people use cloud storage for their security cameras  , or that they select cameras offering both local and cloud options (such as our current top-two indoor camera picks). Although local storage is usually cheap (just the cost of the memory card), and in going local you don’t have to worry about who might potentially view your footage, there are a few specific reasons we recommend *only those cameras that offer some type of cloud service.

*Without the local storage memory card.
What if there is no broadband.

It’s the absolute best option for security

Sounds like BULLSHIT to me.

Cloud plans often include extra features
It often pays to—well, pay. Companies want to get you on the hook for that recurring revenue, so they often include exclusive features and other perks to entice new subscribers and keep existing customers happy.

They don't mention or haven't they heard of a thing called port forwarding or a vpn tunnel.

[Messtechniker]

Your Synology NAS, for example, at home supports many cams.
No need for this cloud stuff. Moreover, you can find VLC
network links for internally networked cams on the internet.
Something the cams sellers will not tell you. The catch in many cases:
finding the precise camera designation.

[coppice]

It’s the absolute best option for security

Sounds like BULLSHIT to me.

Not necessarily. When someone says a thing is good for security, the key question to ask is "whose security?". Generally its not your security they are concerned about, but they will be pushing an idea that benefits someone's security, if its merely their own job security.

[Bicurico]

Everything in life has a price.

Do I like to have my data on OneDrive with Microsoft having potentially access to it? Even, worse, with hackers having potenial access to it? Hell no!

But: I do appreciate and make full use of the fact that all my data is on OneDrive. This allows me to use multiple computers (mainly 3-4) and have everything syncronized! This allows me to not have to bring my computer with me and instead I sit on any of the multiple computers and have all the data with me.

So, putting everything on a balance: I would say the risk/benefit tends to the benefit side.

I know I could have my own private cloud and achieve the same. But how much work would that be?

And, mainly, would it really be more secure? I seriously doubt it.

[CatalinaWOW]

Making your own cloud is probably not more secure, but you have more control over end of life, more control over how it works and probably more control over costs.  Onedrive might still be the right choice, but security isn't the only possible wart.

[madires]

OneDrive (and also other cloud storage services) comes with a non-obvious risk, i.e. a locked account. When MS's AI deems a file to be nasty content your MS account is locked and you lose access to everything associated with that account. There are already several reports of such cases (caused by false positives) and a quite cumbersome process to unlock the account.