From whence the hate?

January 9, 2012 8:07 AM

I've seen a lot of people (mostly those involved with enterprise-level computing in some way) deride the term 'cloud computing', or just the notion of 'the cloud' in general. What's the source of this animosity, and why is it to widespread?
posted by anaximander to computers & internet (25 answers total)

Because any time someone else has your data, you lose control over it.

They could lose it.
They could sell it.
They could hand it off to foreign governments.

It's a neat concept - the idea of your data just being available everywhere. Until you think too deeply on the implication of that.
posted by Pogo_Fuzzybutt at 8:11 AM on January 9, 2012 [8 favorites]

Today's buzzword... you get sick of hearing about it from the misinformed. There're legitimate concerns about its feasibility, too, but I'm not the guy to lay that out for you
posted by MangyCarface at 8:11 AM on January 9, 2012 [4 favorites]

Mostly because it's the latest popular and fuzzily-defined buzz phrase slapped on to common techniques which have existed for a long time: a decade ago cloud computing was called the "thin client," basically.

Personally I have not much problem with the concept per se, just with seeing the term applied all over the place by people who don't really understand what it means just because it's the current flavor of the month. (This month's cloud computing is last month's gamification is the previous month's social computing is etc etc etc.)
posted by ook at 8:18 AM on January 9, 2012 [2 favorites]

I think a big part of the problem is that it's so over-hyped and the definition is so blurred that it's an almost-meaningless buzzword -- sort of like "green" products. This Dilbert comic is really kind of poignant. And while Microsoft is probably doing some cool stuff with cloud computing, their commercials linking MS Office to "the cloud" don't help lend any clarity to what "the cloud" is.

Of course, Pogo_Fuzzybutt's point about handing control over to third parties is also true.
posted by fogster at 8:22 AM on January 9, 2012

Mostly because it's the latest popular and fuzzily-defined buzz phrase slapped on to common techniques which have existed for a long time: a decade ago cloud computing was called the "thin client," basically.,

Ah, that was sort of the feeling I got from reading a few sarcastically-worded comments on articles about some new 'cloud computing' initiatives.

Thanks for the replies! The cloud-hate makes a lot more sense now.
posted by anaximander at 8:23 AM on January 9, 2012

As an IT professional I can see the value for some smaller companies who can't invest in off-site data centers. But Holy Jeebus on a pogo stick are you Frakkin nuts? Using the Cloud is no way to have a valid business continuity plan, no way to guarantee safety and security of data, no way to handle business. I'm sure once the SEC and other Federal and State agencies wake up there will be numerous questions about properly securing personal information, long term storage, etc.

As an example, how many of you think All your cloud photos are still there? What about that bunch that you posted to OFOTO? Companies go out of business all the time. Even Google is not too big to fail..( witness Google +).

FWIW if you are in a job interview and the manager says"Oh yeah, we're moving everything to the Cloud". No matter how ridiculous you think it is you should nod your head and agree ever so enthusiastically.
posted by Gungho at 8:30 AM on January 9, 2012 [2 favorites]

Because nobody can tell you what the cloud is. For many people the cloud is something you use not something you build on. I've seen presentations that equated the cloud with "using websites on the internet." Revolutionary, I'm sure. At best it's about elastic demand based computing and time sharing for huge scale, but typically the enterprise just needs a lot of disk storage, not gigantic parallel computing clusters.

Because we've been building the cloud for years, it just never had a fancy name. The enterprise has plenty of cloud technology. High availability databases, SAN storage, virtual machine clusters, load balanced webservers, etc.

Because the cloud is seen as an alternative to the local IT dysfunctions. The cloud promise is about fluid computing and the elimination of downtime, and yet there have been multiple global outages of EC2.

Because the apps we run right now charge us per core. A thousand CPUs at your disposal isn't so handy when you don't have a thousand licenses.

Because law and enterprise policy prevents us from using the cloud anyways. We can't just migrate customer data from NA to EU. And the clouds charge us extra to guarantee it. So then the CIO asks about private clouds, and in reference to point 2, we have to ask how that's different than the status quo.

Because even if we did save money and get the promised perfect uptime, we'll have to upgrade our internet backbone to support all the traffic going out over the internet. All that email bouncing between employees suddenly runs out over the internet and back.
posted by pwnguin at 8:31 AM on January 9, 2012 [1 favorite]

The over-use of the term is the main reason now.

But I have proudly been a cloud-derider since I first heard the phrase. Especially the term "internal cloud". What is that? A bank of servers in a company data centre running shared applications? Umm.... we've been doing that for decades. Previous terms for the same included application cluster, server farm, and heck, even mainframe.
posted by Diag at 8:34 AM on January 9, 2012

Stupid buzz words.
posted by radioamy at 8:59 AM on January 9, 2012

I think there are two very different set of reasons :

- this is the latest buzzword: everyone uses it for different meaning and it is seen as the solution to "all enterprise IT problems" © - a lot of existing hosting practices are being rebranded as "cloud" for example, but are not really new.

- the computing utility model threatens existing IT. The first implementations were really poor (remember Network Computers ?). The current one are crummy at best. But they follow a path that many disruptive innovations take. A typical reaction by incumbents is to deride such technologies, usually until it's too late.


Note that I'm not saying the cloud is disruptive - just that it follows a disruption path - execution from the "cloud" providers will determine wether it's succesful or not. But it raises already some interesting questions, such as "What is software ?"
also : enterprise IT problems are made of people

posted by motdiem2 at 9:02 AM on January 9, 2012 [1 favorite]

The irritation that I have with it is that it is ill-defined so anyone "can be cloud" and lots of people only see the upsides of the approach without considering any of the downsides.
posted by mmascolino at 9:04 AM on January 9, 2012 [1 favorite]

What they said are good examples, but the big one is "not understood". I mean, there's lots of almost angry misinfo upthread. Example - clustered servers aren't "the cloud" but would run in one. EC2 issues weren't global, but limited to a single zone (US east).

"Cloud" is a shorthand for several different methods of moving the "back end" of IT off the hands of 'business'. Wikipedia has a fairly good overview but a simple way to think of it is - instead of buying a car, you rent one as needed.

If you bought a midsized car, that all you get. If you only have one person using the car twice a week - you still have the same car payment every month. If you need to haul lots of stuff, you either have to make many trips, or you just can't move the items.

Now think of renting - you get only what you need for the job at hand - whether that's a tiny commuter car for the week, a dump truck for the weekend, a passenger bus on a holiday, etc. Pay by what you use. That's the promise anyway..

So why the hate? Imagine if you're a car driver, mechanic, car sales man or some after market car parts manufacturer? Less people owning cars is a scary thing.

and on preview motdiem2 has it in a better wording, although cloud is absolutely disruptive.
posted by anti social order at 9:08 AM on January 9, 2012 [1 favorite]

The concept itself is very hard to define, even by people who know what they're talking about. The real problem (which I suppose is true with most any buzzword) is that people who really truly don't understand it for some reason feel confident using the term with authority.

I.e.,

IT: "OK now it's time to renew our phone contracts, here are the options."

Management: "What about the cloud? Have we considered the cloud?"

IT: "That's... what?"
posted by odinsdream at 9:22 AM on January 9, 2012 [1 favorite]

I'm a former CIO at a Fortune 100 company.

The issue breaks down several ways.

First and foremost, 'cloud computing' has no agreed upon definition other that "not in your data center" so you are pitched "the cloud" for every product. It's an overused buzzword that makes CIOs roll their eyes and check their watches.

Responsibility - As a CIO you are responsible for services being up all the time and if they break, you must fix them. Not having the ability to fix the problem when the problem is "in the cloud" is terrifying. Likely you are not the only customer of the cloud service provider and aren't going to get great information about when things will be fixed. This is the most likely mental stumbling block that will need to be addressed.

Cloud is not same servers in another place - Many vendor offer 'cloud compute and storage" but all they are really offering is to put the exact same servers and storage in a different data center. This doesn't promote change, innovation, or efficiency. It becomes impossible to be agile, promotes legacy systems, and drags thinking into operations instead of improvements. The CFO likes this idea because it shows a savings, but the long term cost to the company is huge as technology becomes a drag and not an asset.

Saas is inflexible - Enterprise apps can be modified to do what ever the internal customer wants. Much politics, drama, and careers surround the constant need to rev software. Saas software usually is one size fits all, minor modification is possible at a cost, but usually cannot be changed in the same ways as internal software.

Latency - Some computing efforts simply need to be done in the same sub-10 millisecond network. It's possible to move some things that require this to the cloud, but in general, cloud computing (AWS, Joyent, etc.) strengths are scalability not compute speed. If you are talking about moving large data like media files or large transaction databases around alot in the hundred gigabyte and terabyte range, you run into big problems getting the data in and out of the company.

Compliance - Sarbanes-Oxley and other compliance issues are made harder by SaaS. SAS 70 stuff helps, but you're getting more boring paperwork putting it outside rather than keeping it in. Heaven help you if you are a global corporation and have to comply with EU stuff and Safe Harbor.

Vaporware - The vast majority of solutions offered are simply Powerpoint presentations of ideas that have not been built. They await the customer dollars to build out the needed resources, so you get hit with all the problems of starting any new service. Even things that have been built, have not usually been scale or process tested in the way you might use it.

Stuff sucks - Bad UI, slowness, lack of features, web server crashes, authentication issues, help desk support, scattered points of contact, etc. I can have all those problems internally, why should I pay you to do it?

I could go on for a while, but the summary is this:

Large companies want software and computing to bend the specific and often unique processes they have, SaaS solutions work best when little is modified and the company bends it's process to use the tools optimally.

Companies hate, _hate_, HATE, to change their internal process. It steps on tradition, turf, power, staff, chains of command, and other 3rd rails of corporate life. Advocating change of business process is career suicide unless you are the CEO or a Consultant.

There's a reason I'm a former CIO. ;)

Forward looking companies need to make the choice to optimize their business process to use better and cheaper cloud tools with full executive support. Unfortunately, you don't find this the case in many places.
posted by Argyle at 9:22 AM on January 9, 2012 [19 favorites]

Because ZOMG WHAT IF YOU STORE YOUR STUFF IN THE CLOUD AND THEN IT RAINS?!!

(but seriously, as others have said, because it's a term often misappropriated by management types who don't know what they're talking about)
posted by qxntpqbbbqxl at 9:23 AM on January 9, 2012

I see the cloud as an MBA solution for hiring systems people, who are generally considered to be a cost center. Where it has follow-on effects that may not be beneficial to the company, they are explained as adjustment periods.
posted by rhizome at 9:26 AM on January 9, 2012

I think the points everyone has brought up above about "the Cloud" in specific are good, but I think this is just a particular example of a very general pattern in software development and information technology, that goes something like this:

1. Some people are struggling with a problem in enterprise-scale IT. Since enterprise-scale IT is essentially always connecting some sort of form-like interface to some collection of database-like data sources with various layers of cruft, brokenness, and noise in between, the problems are always very similar.
2. Somebody is working on a new project. He decides to go in a slightly different tack from usual. Instead of building a bunch of objects to proxy access to his data through, he uses a bunch of plain files. He finds that he cranks through his project quickly, more quickly than he expected, and maintenance is easier, at least for now.
3. He goes to some sort of meetup, or is on a mailing list, or he writes a blog post about his success with using Files-Over-Objects. Other people respond, and they start referring to this approach as "FOO."
4. A couple other people try it, they ditch their ORM layers and their RDBMSs. Some of these people are young, at least in industry terms, and they think this is a really novel approach, so they get disproportionately excited about their "breakthrough" and start talking about it whenever they get a chance: blogs, mailing lists, conferences, etc. They don't remember that people used to do something like this in 80s, and it worked OK for this, but it had integration problems when you needed to do that, etc. They write a few libraries to "simplify" the building of FOO-oriented applications.
5. Now there's a bit of a hype storm brewing. This is the first wobble in the coming gyre. Maybe O'Reilly Radar posts about it. The pressure is dropping. The wind is picking up. Blog posts from outside the original group of people start appearing, on more general-interest websites: "Should you try FOO?", "A FOO Tutorial in 10 minutes", "One-click FOO installer." The key sign at this point is that it becomes easy to use the new thing on Windows. Once someone has gone to the trouble of making something work easily on Windows, the tipping point is at hand.
6. Switching metaphors entirely, the charlatans smell blood in the water. The less-informed megasites and trade magazines start writing articles about "The FOO Revolution." They organize conferences around FOO. They start publishing trade mags called "FOO Strategies Today." Someone takes one of the open-source FOO libraries and starts selling support contracts for it. Money is being made.
7. Executives who don't know their ass from their elbow read about how a FOO initiative saved MegaTron Corporation 40% in year-over-year MPM (Meaningless Performance Metric) in a shitty magazine on a plane. Neither the quoted MegaTron executive, the writer of the magazine piece, nor the reader of the article understand the first thing about software development. Actually, none of them are even remotely interested in it. The writer wanted to write about other stuff, but has this job temporarily because its hard to get a writing job, the execs were both MBAs who view this as widget production. Who cares about the widgets? It is completely irrelevant to them.
8. As soon as the plane lands, the executive starts emailing their staff: "What is our FOO strategy? Are we moving ahead with the FOO project? Kill the 90% done traditional project and replace it with a FOO-oriented project or we're basically throwing money away!" The executive doesn't know anything about software, but remembers some mumbo-jumbo from business school about T and comparative advantage and IRRs.
9. The engineers and managers on the project groan. They point out how FOO is really not that different from Basic Access Read-only (BAR) which they had been using, and that it just moves complexity from up-front development into later cross-project integration. There's no free lunch, no silver bullet. They tell the executive this. The executive is incapable of telling the difference between a computer that is broken, and a monitor that has been turned off.
10. Now, the team has to call a bunch of vendors and say, "Hey, do you have a FOO-compliant version of your system? We were on the BAR version, but we're moving to FOO." The vendor sales people panic. "Yes, of course we do! It's shipping this quarter!" Then they tell the project managers, who say "Wait, a FOO version of our system? That makes no sense! We make disks! Making a FOO version of our system is like making a fragrant version of purple. Those terms just…don't go together." The salespeople say "I don't care, the customer wants FOO on the box. We need a bullet point on our 1-sheet that says "Full FOO Compliance or we're going to blow the quarter!"
11. Management says, "Shit, not blow the quarter! My bonus is based almost entirely on our stock price at the end of this quarter, and I really want to buy a beach house! Fuck it, do what the sales people say! Our stock price will definitely get beat up if we blow the quarter!
12. So the PMs talk to the engineers and come up with some insane way to argue that their system is FOO compliant. To do this, they ignore working on real bugs, performance issues, and long-standing problems, to chase FOO compliance, so they can sell it to a client who literally has only a hazy idea of what FOO compliance even is or what they want. (What will really happen is the vendor will "sell" the client software, but for every dollar of software sold, they will sell six dollars worth of "service", wherein they actually develop the software the client needs, possibly.
13. Now, every vendor in the world is saying they have FOO in their new version. Every executive is demanding to buy FOO. Neither the vendors nor the IT purchasing execs really understand what the original purpose of FOO was from way up at the top of this story. Basically anything and everything is getting rebranded as FOO. This is the year when the FOO Conference has its peak attendance and sells out all its tradeshow floorspace. At this point, the word FOO has been nearly bleached of meaning. A few startups that claim to build "FOO lifecycle management technology suites" are bought for eye-popping sums by Oracle, SAP, etc., despite the fact that they have been basically building custom contractware for a few large accounts and calling it a product.
14. A few more months go by. Now, projects that were switched to FOO are running late, over budget, under featured, etc. Like every other IT project since the beginning of the world. Even at the executive level the realization that FOO doesn't magically solve problems begins to take hold. The words of the greybeards get some attention: "FOO is basically a rehash of Basic Allocation Zones, the BAZ strategy we had back in the '80s, and just like in the 80s, it worked well in some cases, but you still need to worry about X, Y, and Z. " On top of this, many of the FOO projects have absolutely nothing to do with the original FOO ethos, because of FOO meaning dilution, so these engineers are struggling with the same problems as ever, only now they are called FOO problems.
15. Now, workaday engineers are familiar with the fact that the meaning of the term FOO has been fully stripped out, its just an empty marketing term. They're also aware that there are just as many borked FOO projects as non-FOO projects, and while even the original understanding of the FOO technique might be good for some wins in some cases, it also has some costs that come up and might outweigh its advantages in other cases.
16. Consequently, whenever someone says, "Hey! Have you seen our new FOO-centric IDE? ITS GREAT!" the engineers, PMs, and technical folk take this as a sign of cluelessness. It's like a reverse-shiboleth.
17. Meanwhile, some kid somewhere is hacking a project that replaces all these plain files with a database that the application code accesses through an object layer. He calls it Objects-Over-Files…

And that is how the IT industry, bedrock of our economy, last big win in growing our gdp-per-worked-hour, works!
posted by jeb at 10:55 AM on January 9, 2012 [58 favorites]

Wow. I think we're done here.
posted by odinsdream at 11:08 AM on January 9, 2012 [2 favorites]

I don't know about cloud computing specifically, but enterprise IT managers dislike of trendy-concepts-du-jour usually stems from two root causes. a) Trendy concepts are often popular with top management types who don't really understand them or all the issues involved, but are always looking for some panacea to solve their IT problems. b) Like everyone else IT managers can be resistant to upheavals in how they do things, and want to protect their turf.

It's probably ever been thus since the days when mainframes ruled and IT managers weren't keen on those newfangled minicomputers popping up all over the place, outside of their control.
posted by philipy at 11:16 AM on January 9, 2012

I'm a cloud-hater for many of the reasons detailed here, but sometimes the cloud is dismissed by people in IT who stand to lose out. There are many organizations who are too small to justify the in house expertise to manage their systems and data. Unless they are growing rapidly they are often wise to go leaner on staff in favor of a cloud provider. The well deserved criticism is that cloud computing is being pitched and bought in environments where it doesn't add as much value and for some reason it has become considered a safe sale for all parties in spite of its weaknesses.
posted by dgran at 11:22 AM on January 9, 2012 [2 favorites]

It's yet another buzzword for one of these things that's the business-world equivalent of sex when you're a high school student: As far as you can tell, everyone else is doing it more and better. And even the folks who are doing it are flailing around and are convinced that other folks are doing it better and they don't want to let on.

At its basic point, "Cloud" just means "Someone else's hard drives".
posted by rmd1023 at 1:25 PM on January 9, 2012 [2 favorites]

Nthing what most people say, but there are a number of cloud based services that work very well. Developers roll their eyes when someone in management mentions the cloud, but it doesn't stop them using cloud services like dropbox, github, Facebook and basecamp. You need to ignore the usage of the word and concentrate on the service supplied.
posted by seanyboy at 2:54 PM on January 9, 2012

I'm just sick as hell of hearing people append the word "cloud" to things to make their product sound cool and modern. Especially since "cloud" could mean anything from "distributed redundantly over many datacenters" to "IT'S ON THE INTERNET!!!! IT'S A WEBSITE!!!"
posted by Afroblanco at 3:31 PM on January 9, 2012

Nobody can say it any better than jeb just did.

After working in the industry long enough the hype-storm process becomes second nature. I have watched the same cycle of magic, boom, hype, bust, reality happen with OOP, Java, XML, the web itself, thin clients, fat clients, "push technology" (never did figure out what that meant), agile development, automated refactoring, test-driven development, and of course cloud computing. I'm sure there are plenty more I'm just not remembering at the moment.

I'll know the next hype-storm is on its way when I hear a lot of people excitedly talking about an amazingly powerful architecture change which is guaranteed to speed up development by 10x and make everything I know irrelevant. When I look into it, I'll find that there's no there, there - it's a minor, incremental advance that barely deserves a name, much less a movement - but when enough people crowd on board quickly enough, absolutely anything can take off and become a big deal through simple social mechanics.
posted by Mars Saxman at 4:20 PM on January 9, 2012

I want to be like jeb when I grow up. That answer rocks.
posted by dgran at 9:50 AM on January 10, 2012 [1 favorite]