How Increasing Cloud Reliance Affects IT Jobs

snydeq writes "Kevin Fogarty takes a look at how the rise of cloud computing will impact IT jobs, outlining which roles stand to gain prominence in the years to come, and which roles will suffer as organizations extend their commitments to the cloud. 'Ultimately the bulk of IT could look more like a projects office than the way it looks now, when most of the hands-on work is done inside. It probably won't be a total transformation, but moving into cloud, there will be more of that and less DIY.'"

Comments

Not what I signed up for

[tverbeek]

Great, I was hoping they'd get around to removing the last bits of enjoyment from my career before I retire. They're even ahead of schedule!

Re:Not what I signed up for

[Concerned Onlooker]

You were lucky. At least you got to have some enjoyable bits. Nowadays it's all about system engineering and learning to use MagicDraw.

Re:Not what I signed up for

[tubs]

Don't worry, in a few years a new company will come along, with a great new product that will allow you to cheaply pull the information back into your organisation, handing power back to users, distributing access and design.

This will then be followed by a period of great excitement, with some people making themselves rich, but then that company will become large and bloated, creating more and more bloated systems, and then we'll be sending our information back out to a "central" system.

Re:Not what I signed up for

[Gideon Wells]

At least you got to enjoy it for a while. All my professors during my final three weeks of school at college started coming out and admitting to my sub-major group that we had spent four years studying for a profession that was en route to be obsolete by 2012 as we learned it.

Dear god I hate my college. On the bright side, IT was my alternative career choice. It looks like either way I was heading for the same route.

Re:Not what I signed up for

[tverbeek]

In all seriousness, if you're just getting out of college, that means you have plenty of time to change direction, start over if necessary, and find some other profession that you will enjoy (at least while that one lasts; some are perennial). Not so easy when you're a quarter century into one.

Re:Not what I signed up for

[jimicus]

At least you got to enjoy it for a while. All my professors during my final three weeks of school at college started coming out and admitting to my sub-major group that we had spent four years studying for a profession that was en route to be obsolete by 2012 as we learned it.

Dear god I hate my college.

Count yourself very lucky that they told you that. Most people come out of college with no real experience to speak of and spend the first 6-12 months of employment learning all about the things they don't teach you in college. And that's the sort of thing you seldom get taught.

Truth is, IT management, maintenance and support is well on its way to being a commodity - which means the only way to make it work is "pile it high and sell it cheap". For many small businesses IT already is a commodity - they set up some sort of support agreement with a local company and if they're not happy, move onto the next one. The USPs you can think of to attract and retain customers (if you're an outsourced service provider) or keep your job (if you're in-house) are rapidly becoming less and less interesting.

Re:Not what I signed up for

[Anonymous Coward]

This is why I no longer deal with computers, IT, or anything remotely related to networks anymore. The stupid managers , asskissers, and teeny boppers think this is something new, just like everything else for the last 4 decades. No it's the same old idea but because of more bandwidth and speed across networks we are able to move more and more of the servers offsite. Big fucking deal. It's client/server computing and if you think it's anything different you are an idiot. FYI facebarf and twatting are not new ideas either.

And when the cloud goes down.

[Maxo-Texas]

Your business is dead in the water.

If that only happens 2 days a year, you just factor that as a cost unless there is some critical reason you must remain up (hospital).

Also, it becomes difficult to differentiate your business from others.

As jobs get completely slaughtered something has to give. Shorter work weeks or civil unrest.

Re:And when the cloud goes down.

[TooMuchToDo]

Excellent point. Netflix has moved almost all of the infrastructure into Amazon AWS. If you want to buy Netflix, I'm assuming they'd hand you the keys to the AWS login and the contracts they have in place with media groups (yes, the distribution centers are a different story, go with it for the example).

What sort of value does your business have if someone else is running all the infrastructure behind it? And that infrastructure company sells resources cheap to anyone else in the world? You could build a Netflix, a Dropbox, a DNSMadeEasy/DynDNS/UltraDNS, etc., as long as you have a team with the time to build it.

Exciting, interesting, and worrisome at the same time. As OP said, short work weeks or civil unrest. May you live in interesting times.

TL;DR Massive, cheap infrastructure-on-demand drags down valuation of businesses that run on it. Win for consumers (hopefully), lose for businesses (not so bad).

Re:And when the cloud goes down.

[lpp]

Also, it becomes difficult to differentiate your business from others.

Right now, the differentiator between you and me should not be how we store our data. Whether my data resides on a server in my office or in a databank with some outfit I can only access through my ISP, what decides it for our potential clients should be the quality of service we offer respectively, or even the types of service we offer.

Re:And when the cloud goes down.

[Anonymous Coward]

I am on board with shorter work weeks. A 4 day work seems adequate for most professions. Heck we could even have 4 9 hour days instead of 5 8 hour days if shaving 20% off the work week seems too big at first. I would think most people would be happier even if each workday is a bit longer.

Re:And when the cloud goes down.

[hawguy]

You could build a Netflix, a Dropbox, a DNSMadeEasy/DynDNS/UltraDNS, etc., as long as you have a team with the time to build it.

Isn't that true regardless of whether they host the apps in the cloud or not? It's the software that's hard to replicate. While your team tries to duplicate Netflix's software, your hardware team could be racking servers in the datacenter(s).

Netflix probably pays $200M/year or more in bandwidth costs - the cost of hardware to pump out the bandwidth pales in comparison to the bandwidth itself.

Perhaps you can save some money by going to the cloud, but at the scale of a company like Netflix, I'd be surprised if the AWS is significantly cheaper than hosting their own hardware. Amazon can't rent out a thousand virtual servers for less than the cost to run a thousand virtual servers. If a company needs 10's or hundreds of servers then I can see how Amazon could be cheaper, but once you get up to Netflix's scale where their needs can consume entire datacenters, then the economics become less clear.

Re:And when the cloud goes down.

[Belial6]

The difference is that a certain portion of the IT staff - the ones who train users on Microsoft's latest and greatest office suite - won't be required any more.

This is the kind of idea that will cause HUGE problems for businesses that buy into the cloud hype. How in the universe do you think that having the appli8cation hosted off site, will reduced in any way the need to train users on the software? That is just a bizarre statement. The only difference when it comes to training between an app hosted on site and one hosted on the 'cloud' is that the one on the cloud may get upgraded whether your trainers are ready to make the move or not. On site upgrades means that you can schedule your application changes, and can roll them out in a staggered fashion.

Re:And when the cloud goes down.

[yarnosh]

If that only happens 2 days a year, you just factor that as a cost unless there is some critical reason you must remain up (hospital).

Not sure I see something like a hospital move to the "cloud," but stranger things can happen. I mean, I'm pretty sure they're smart enough to realize that they can't bet everything on their internet connection. For most everyone else.. meh. It isn't like in-house systems never go down. You might have an awesome team who keeps things running like a well oiled machine, but I hardly think that's much of a candidate for moving to cloud anyway.

Also, it becomes difficult to differentiate your business from others.

If you're relying on tools/software to "differentiate" yourself from other businesses, you're doing it wrong. The kinds of things you're going to be moving offsite into hosted services in the cloud are usually very generic things anyway. Stuff that most businesses just need to get done as cheaply and as efficiently as possible, and it really doesn't have a whole lot of direct impact on customers.

As jobs get completely slaughtered something has to give. Shorter work weeks or civil unrest.

Meh, Somebody has to run those data centers and cloud services you know. Computers and automation have always threatened to put people out of work but it never seems to actually work out that way. You just have to know how to follow the jobs. I mean, you can sit around bitching and complaining that your job running your company's mail server is gone because they moved everything to Gmail or you could apply at Google. Or start your own cloud service. Or maybe you were just another clueless small potatoes sysadmin who shouldn't have been working with computers in the first place...

Re:And when the cloud goes down.

[yarnosh]

What sort of value does your business have if someone else is running all the infrastructure behind it?

I'm not sure you really "get" how services like Amazon AWS/S3 work. All they do is manage the physical infrastructure at the lowest level. You still have to write your software to use all of it and tie it together. That's where the real value lies (in terms infrastructure). Making good use of Amazon is no trivial matter.

You could build a Netflix, a Dropbox, a DNSMadeEasy/DynDNS/UltraDNS, etc., as long as you have a team with the time to build it.

Absolutely. And that's what makes services like Amazon and Terramark so compelling. The thing is, not everyone can't be successful at it. Anyone could go out and start a website for almost nothing. Getting people to visit it is another story....

TL;DR Massive, cheap infrastructure-on-demand drags down valuation of businesses that run on it. Win for consumers (hopefully), lose for businesses (not so bad).

WTF are you talking about? How could cheap, available infrastructure possibly drag down valuation of a business? If you have X $$$ to invest in your tech business, you can either dump it into buying and maintaining your own hardware or you can put it into software to tie mature cloud services together.Either way, the value is there (assuming you invested wisely).

Re:And when the cloud goes down.

[trevelyon]

I see more than a few problems with this article not the least being the base assumption that cloud computing will continue to grow at the pace it has so far. Cloud computing has the following major issues that still need to be addressed:

1. Reliability - Cloud services still can not match well managed in-house IT for dependability IMO
2. SLAs - Anyone that has actually negotiated and enforced SLAs knows what a joke they are. Unless you get the entire month's service cost for free when they provider misses the mark (and no-one does this) the penalties are so insignificant as to make even having an SLA pointless. By the time the outage hits the level where you will feel the actual discount your entire business could be gone completely (i.e. you've been down a week solid).
3. Security - The more you trust unknown providers with unknown security implementations and share hardware with unknown entities the less control you will have over security issue. As more and more breaches occur untrustworthy organisations will be punished by their customers for their lax security.
4. Data Lock In - reliance on many SaaS providers will sooner or later lead to you being able to not always get your data in the form you wish. It's the cloud version of Quickbooks, great while you are happy but just try to migrate all that data to something else. Note: this is not true of all SaaS providers but is of some, sooner or later without set standards on what to look for one of the lock in providers will be chosen.

The other main problems with the articles are:
1. Assumes hardware swap or VM reimage will fix most problems - Not sure where they got this idea but someone still needs to configure the software services for your specific tasks. A general VM is not the same as the customized config you use for your implementation, cloud or not.
2. Assumes SaaS works on an enterprise scale - My experience is that for most things it simply does not. SaaS offers very good value for the SMB market but for most enterprises it has major issues that stop it from competing effectively with in-house applications. The largest one IMO being integration with other in-house systems and data sharing. Almost all SaaS offerings fall short in these areas and can not compete with a well run internal IT dept over even the mid-term. Sure there are short term gains in getting something quick into use but move down the timeline a few years and you find you rapidly end up with a spiderwork of hacks trying to tie systems together, systems you have very limited control over. In the end it pushes more responsibility to the "glue systems" changing your IT to a very reactionary structure. Good luck should one of your SaaS providers go out of business or change their API drastically.
I am not saying there is no place in the enterprise for cloud computing, there quite obviously is. What I am saying is it is one more tool a well run enterprise IT dept. should use but use cautiously. It's a much more ready fit for the SMB market that normally does not have the resources to integrate it's systems and is used to running with a slew of unconnected data. Enterprises need not only the ability to access their data but a measure of flexibility in how it is structured and accessed. Time will tell what the true cost but I do envision a similar situation to outsourcing of helpdesks, we may see many apps sent to the cloud return to in-house systems after they do not meet the needs of the enterprise.

Re:And when the cloud goes down.

[TheRaven64]

Not the grandparent, but I would imagine that the company providing the application would provide training if you buy the expensive package. If you buy the cheap package, then your users just have to work it out for themselves.

Re:And when the cloud goes down.

[hoppo]

At that scale, though, Netflix probably negotiated a pretty sweet deal with Amazon. A flagship customer like Netflix really overcomes a lot of the objections customers may have. It's a pretty powerful answer to any company who questions whether or not Amazon can operate on an enterprise scale. Odds are, Amazon is sacrificing its profit margins, possibly even eating a loss, to host Netflix's operations.

Re:And when the cloud goes down.

[andy1307]

Your business is dead in the water.

Yes..because that sort of thing never happens when you host the services in-house.

Re:And when the cloud goes down.

[Maxo-Texas]

At my company they have solved this problem by no longer training people. In just about anything.

We went LIVE on SA FRIKKIN P with NO training. How can we demand 5 years experience of new people?

We will get a small training course sometime later this fall.

Re:And when the cloud goes down.

[Colde]

But if your reliability depends on the same cloud as your competitor, the reliability is basically the same.

Re:And when the cloud goes down.

[just_another_sean]

I generally have to work 9 or more hours in a day anyway. Sure, some of it is spent on /. but there are more then enough 10+ hour, "must get shit done" days to make up for the slow ones. An extra day off in a week would make me more then willing to make the other four 10 hour days.

Re:And when the cloud goes down.

[TooMuchToDo]

First, yes, I'm well aware of how Amazon's AWS infrastructure works. If you're a business, you're making money either via a) labor (hiring people; their output) b) rent (licensing land, IP, computing time, etc).

If you're a business that makes money where the majority of the work is done by your computing equipment (Netflix, etc), the business is mostly that equipment. Sure, you've got licensing. Sure, you've got distribution centers for the physical discs. But if your business is mostly that software, and the hardware side is elastic and fungible, while also being cheap and easily obtainable by anyone, your business is some sprinkles on top of a cake anyone can buy a piece of.

A fairly large component of a business is barrier to entry. Can you compete with Walmart? No, of course not. Enormous capital outlays would be required.

The capital outlay is time to build software systems like Netflix that run on Amazon. You may argue "Who is going to waste their time with that?" Easy. The 4-6 billion people not in the first world (Brazil, Russia, India, China, as well as a lot of eastern Europe) whose time is relatively cheap. You just need a credit card and an AWS account. Low barrier to entry = low business value.

Don't be shocked as not only the cost of online services that anyone can develop drops, but also the available profit margin of said services.

http://en.wikipedia.org/wiki/Porter_five_forces_analysis

Re:And when the cloud goes down.

[TooMuchToDo]

I think it also takes a bit away from Netflix by running on Amazon. You look at Facebook, Google, etc. and you'd never dream about running that stack on Amazon's infrastructure. But Netflix? How special can what they do be if they moved everything into Amazon?

Having your own gear has an air of "We're doing something different, something hard" when you're a huge tech company. Running on Amazon says "We're a commodity, and it's not hard to replicate what we're doing".

Re:And when the cloud goes down.

[ultranova]

As jobs get completely slaughtered something has to give. Shorter work weeks or civil unrest.

Or more likely longer workdays with less pay - after all, your negotiating position becomes worse and worse - after which you'll get taxed more to hire more police to deal with security problems. Any attempt to unionize and deal with the problem will fail because then someone else might get more than they deserve, and in any case you are better than average and get to dictate your own pay any day now. And besides, it would send all the jobs to India, right?

Re:And when the cloud goes down.

[Maxo-Texas]

With robotics and automation you could be looking at extended 20% unemployment with no benefits.

Hungry people tend to get violent and take desperate actions. You can't imprison 10% of your population. What's coming up in the next 10 - 20 years is a change on par with the industrial revolution-- only there are not enough jobs for humans on the other side of that change.

There will be more smart, more educated, more dumb manual labor that is required if we continue working the hours we have been.

The catch is, it's great if you have 20 employees doing the work of 1000. It falls apart when no one else hires those 980 people. Being productive doesn't help if there is no market for your product.

That's the nub of the recession we are dealing with now. And a lot of it is from the fact that companies are working people 60 hours a week without over time using the "exempt" status. Start giving time and a half for anyone who isn't managing other people and unemployment drops by 2% over night.

Re:And when the cloud goes down.

[AK Marc]

The issue is that the people who are the only IT person at a small to medium sized company are unqualified for a job at the hosting company. Maybe they had a CCNA and some other certifications or equivelent experience. But the hosting company needs someone with BGP experience, and a company with just one IT guy wouldn't be messing with that, and if they are wouldn't have someone messing with it more than once a year. But the Hosting company is bringing on people all the time. Or the small company guy does some casual work with Linux or Windows, but isn't a master of them. Again, same problem. And his ability to fix an old HP printer is worthless in the hosting company.

In theory, it should just move jobs, but the problem is that the level of specialization of those jobs is vastly different than those its replacing such that few could get a job at the hosting company. Also, the hosting company would have fewer jobs, otherwise they wouldn't be able to compete on price, so there won't be a complete movement of jobs, some would necessarily be lost.

Re:And when the cloud goes down.

[yarnosh]

If you're a business that makes money where the majority of the work is done by your computing equipment (Netflix, etc), the business is mostly that equipment.

You couldn't be more wrong. The value of the physical hardware, were you to liquidate it, is generally a relatively small part of most businesses unless you are Amazon or Terramark and renting out hardware IS you business. ANd if you are one of those companies for whom the hardware is that much a part of your total value, you probably wouldn't profit from renting it from someone else, so you won't be using Amazon anyway. Cloud services like Amazon AWS are especially for companies for whom the hardware is a small part of the business.

Sure, you've got licensing. Sure, you've got distribution centers for the physical discs. But if your business is mostly that software, and the hardware side is elastic and fungible, while also being cheap and easily obtainable by anyone, your business is some sprinkles on top of a cake anyone can buy a piece of.

So so wrong. If you think anyone can develop the next Netflix killer by sprinkling some sugar on Amazon AWS, have fun trying.

The capital outlay is time to build software systems like Netflix that run on Amazon. You may argue "Who is going to waste their time with that?" Easy. The 4-6 billion people not in the first world (Brazil, Russia, India, China, as well as a lot of eastern Europe) whose time is relatively cheap. You just need a credit card and an AWS account. Low barrier to entry = low business value.

But that low barrier to entry doesn't seem to be helping any of those countries compete directly with first world corporations.

Don't be shocked as not only the cost of online services that anyone can develop drops, but also the available profit margin of said services.

The cost of simply getting a site out there has been negligible (not counting developer time) for YEARS now. The better part of a decade, even. It is thinking like yours that caused the internet bubble of the early 2000's. People saw how easy it was to put up a site and didn't realize that monetizing it is damned hard. I might go so far as to suggest that profiting from a website is getting more difficult. Used to be you could make OK money with a small site and a few ads. Now advertising margins are tiny. You need to be pretty big to make a reasonable amount of money from advertising now. Ok, maybe those profits are going to be sufficient for people in third world countries, but the vast majority of them will remain bottom feeders, not the next Netflix.

Re:And when the cloud goes down.

[TooMuchToDo]

So so wrong. If you think anyone can develop the next Netflix killer by sprinkling some sugar on Amazon AWS, have fun trying.

Really? Seriously? I agree that their recommendation engine is extremely valuable. But the rest of it? Oh so hard to rip content, shove it into S3, and use EC2 instances to serve the content back out to registered users.

This is what we call "race to the bottom". Whomever is willing to take the least amount of profit is going to win.

Re:And when the cloud goes down.

[yarnosh]

Really? Seriously? I agree that their recommendation engine is extremely valuable. But the rest of it? Oh so hard to rip content, shove it into S3, and use EC2 instances to serve the content back out to registered users.

Oh please, anyone willing to risk using your fly-by-night service is just going to torrent the shit for free. And they'll probably get more content that way. Hell, I pay for Netflix and still have to go to torrent sites to get my 30 Rock fix (no cable TV). You'd be a bottom feeder, at best.

You have it so backwards. People accepting the least amount of profit is what caused the big internet boom and bust 10 years ago. LIke you, they thought that anyone could make a profit by putting up a website.. any website. Didn't really matter what the website did because the barrier to entry was so small.How wrong they were.

Re:And when the cloud goes down.

[ultranova]

Hungry people tend to get violent and take desperate actions. You can't imprison 10% of your population.

What's stopping you? Aren't most dictatorships basically giant open-air prisons? And aren't unemployed already basically treated as criminals on probation?

The catch is, it's great if you have 20 employees doing the work of 1000. It falls apart when no one else hires those 980 people. Being productive doesn't help if there is no market for your product.

What happens is that the rich barter amongst themselves and live in even more luxury, the lucky 20 serfs get thrown enough scraps to survive (until their jobs are also automated), and the 980 get the choice between starving to death or taking on automated turrets. Or, if they're really lucky, they'll be allowed to farm land owned by the elite, for this which the latter will no doubt consider themselves as great humanitarians.

Automation in a capitalist society means more and more people are removed from the economy entirely. We either move to a new economical system, or accept that the future will be a nightmare for the vast majority of us.

Re:And when the cloud goes down.

[Maxo-Texas]

I couldn't agree more.

What's stopping me? I'm fine. It's my friends who are not. And large concentrations of people in other states. I'm drawing a conclusion based on history.

Fortunately we don't have automated turrets yet. And if we did, they would have points of failure. The end result is a Somalian situation. The rich there don't live in the same kind of luxury and safety that the rich here do.

I think I'll be very old or if lucky- dead- before it all comes to pass. But I do speak and vote towards a better society with shorter working hours, reasonable tariffs on products made in countries which destroy the environment (we shouldn't be allowed to purchase from them to get around our pollution laws).

Re:And when the cloud goes down.

[Anonymous Coward]

Also, what happens to your business when your "cloud" provider goes down for good? Can you get a copy of your data? What about your apps/services there and any data they contain (e.g., e-mail, web sites, and DBs)?

Re:And when the cloud goes down.

[Maxo-Texas]

A related issue we see is that since our people do not get wide exposure or training, there are classes of errors which are hard for them to fix.

We (finally) had an outside expert come in to audit our settings on our hardware and the changes resulted in a roughly 5% improvement in performance without causing any failures. 5% doesn't sounds like much but it's probably a few million in hardware so even $200 an hour for a week is justified. It's a one time thing- but at least have them drop by once a year IMHO.

Re:And when the cloud goes down.

[ultranova]

What's stopping me? I'm fine. It's my friends who are not. And large concentrations of people in other states.

Are your friends and these large concentrations of people in other states rich? If they're not, are they willing to unionize and vote in socialists? Because any attempt to solve any problem in capitalism is socialist, at least from capitalist perspective.

I'm drawing a conclusion based on history.

So am I. Frankly, human history is not very nice.

Fortunately we don't have automated turrets yet.

Yes, we do. And the video is that of a toy; guess what the military has?

And if we did, they would have points of failure. The end result is a Somalian situation. The rich there don't live in the same kind of luxury and safety that the rich here do.

Technology has marched on since Somalia collapsed, and besides, it still collapsed. Also, think about it: you're comparing us to Somalia. Is that not a sign that the situation is dire indeed?

I think I'll be very old or if lucky- dead- before it all comes to pass.

It's already coming to pass. You said yourself the current economic problems are related to this.

But I do speak and vote towards a better society with shorter working hours, reasonable tariffs on products made in countries which destroy the environment (we shouldn't be allowed to purchase from them to get around our pollution laws).

Well, with Germany abandoning nuclear power I think the War on Global Warming is lost - we've already pretty much cut all other kinds of pollution well below what nature can take. We'll just have to adapt to the changes in our environment. But hey, with luck the adaptability required for that will draw out the transition to a fully automated economy so much the masses will awaken and demand their rights before its too late.

One can hope, at least.

retrain as a lawyer

[Hognoxious]

Sod IT, go to law school. When it's all up in the cloud and the cloud breaks there'll be a killing to be made.

Re:retrain as a lawyer

[c0lo]

Sod IT, go to law school. When it's all up in the cloud and the cloud breaks there'll be a killing to be made.

If "killing" is what you are after, then yes.
Speaking for myself, I'd prefer to retrain as (for example) a driving instructor: at least this job won't be outsourced anytime soon and in the spare time I can write some open-source. Remember the end of the "Office space"?

Re:retrain as a lawyer

[Lord Juan]

And while that happens you can always work in a patent or copyright troll firm.

Honestly, I can't even comprehend how is anything moving anymore in the US. I also can't see why would anyone who isn't already in a good IT position want to enter the field nowadays.

Re:retrain as a lawyer

[Lord Juan]

At least until driverless cars become the norm. But yeah, I think you still have quite a few years left of that.

Re:retrain as a lawyer

[Warlord88]

Interestingly, software is affecting law industry big time. Gone are the days when lawyers needed to burn the midnight oil to research relevant cases. With a document managing software with proper tagging features, even you can become a lawyer!.

Re:retrain as a lawyer

[c0lo]

Honestly, I can't even comprehend how is anything moving anymore in the US. I also can't see why would anyone who isn't already in a good IT position want to enter the field nowadays.

To the point that producing software in US will be so devalued that nobody will create anything ... except open-source? I don't know if I like or dislike the idea.

Re:retrain as a lawyer

[xtal]

It's funny, but if you can stomach it it's not a bad idea for an engineer with a lot of industry experience.

Re:retrain as a lawyer

[jimicus]

Don't you believe it. It's a fantastically competitive field to get into, so much so that it's quite common for reputable law schools (and for that matter law firms) to demand every applicant have perfect exam results going right back to their schooldays even if they enter the profession as a mature student many years after they left school.

Re:retrain as a lawyer

[AtlantaSteve]

I'm a 12+ year Java developer, who recently completed a JD at a T2 law school. I was basically bored and unsatisfied in my career. I still love to code, but I've seen pretty much everything there is to see... and I spend 95% of my time in meetings or wrestling with environmental dependencies rather than coding.

However, I've stayed in I.T. regardless, because the grass is NOT greener on the other side. As with anything else in society, the top-5% of lawyers are doing great... but things are miserable for the bottom-95%. It's the worst legal job market in almost a hundred years. It can take a year or two of searching to find a legal job, and the only legal jobs available consist of soul-crushing drudgery (even by I.T. standards). Finally, the average salary for non-top-5% lawyer is about 50% below that of an experienced Java developer (who can always land a new job on a few weeks notice).

I know that the parent comment was played for sarcasm, but don't believe the hype. The legal field sucks much worse than I.T.

Re:retrain as a lawyer

[MoralHazard]

Fuck law school, and feel sorry for the lawyers. You don't know how bad those kids have it, these days.

Nationwide, salaries for starting law school grads have been dropping steadily since 2008. The job market for JDs is utter shit, and will probably stay that way for a while. Much of the classes of 2008-09 are still looking for their first "real" jobs (i.e., requiring a laws degree, not a Starbucks' apron). Check out the numbers:

  * http://www.payscale.com/research/US/Job=Attorney_%2F_Lawyer/Salary

That's right: $45K/year for a job that requires 3 extra years of school after your B.A., and costs you $150,00 in tuition. (And that's IF you land a job.) And while salaries do increase with experience, 95% won't break the $100K/year mark until they've been practicing for 8 or 10 years.

I have a couple of dozen friends who started law school (circa 2001-2009) instead of getting real jobs right away, most due to a combination of:

    1) they saw it as an easy ticket to big money and professional distinction, unlike their recently-earned liberal arts B.A. degrees
    2) they hadn't been thinking in specific terms about their post-graduation career plans, and so they basically had no idea what they wanted to do with their lives
    3) basically any bank would approve full student loans for anybody who got into an accredited law school, at the time.

The people I knew were just a small selection of the wave of under-motivated, unskilled, and barely employable Gen-Y folk who flooded into US law schools. And since the law schools have zero accountability for whether their students get jobs after graduating, there's no direct feedback to stop them from churning out J.D.s into an already over-saturated swamp of 20somethings drowning in unsubsidized student debts (at sky-high, pre-crash interest rates, too).

A few of my friends were the lucky winners who landed the really good "big firm" jobs, with super-high starting salaries--this is maybe 5% of all law school grads, nationwide, in any given class. People who got hired at Skadden, Proskauer, MoFo, etc. were making $150K-$200K, right out of law school. These jobs are reserved for the super-high achievers, the kids who've been busting their asses studying every night/weekend since high school, with straight 4.0 GPAs since before junior high. In their first 5 years, they're each required to bill at least 2,000 hours/year to clients, which really means spending 60-80 hours/week at work (not everything is billable). Standards rise as they advance, and if they want to make partner, they'll have to put in around 80 hours/week for the next 10 years or so--and keep that pace up if they want to stay partners in the firm. Salaries do rise, too: A 5th-year associate at Skadden can pull down $500K/year, in exchange for her 80-100 hours/week, and a partner will bring home multiple millions/year.

Re:retrain as a lawyer

[Anonymous Coward]

If "killing" is what you are after, then yes.
Speaking for myself, I'd prefer to retrain as (for example) a driving instructor: at least this job won't be outsourced anytime soon and in the spare time I can write some open-source. ...

Until cars are driven by computers and you are out of a job again.

Re:retrain as a lawyer

[hoggoth]

Don't count on it: http://www.smartplanet.com/blog/thinking-tech/googles-self-driving-car/5445

Balderdash

[Nefarious Wheel]

"Cloud" exists because the MBA's wanted their own word for Internet.

Re:Balderdash

[cloudmaster]

No, it's because "virtualization" is way too hard to spell.

Re:Balderdash

[c0lo]

No, it's because "virtualization" is way too hard to spell.

that and v13n is hardly a buzz-word.

Re:Balderdash

[mini me]

The Cloud is an abstract interface, not a specific technology. It always has been. Look at some networking documents from years ago and you'll find the cloud present. The cloud services might be implemented using virtualization, but you don't care, because it is just an abstract network that you throw your bits at.

I think it is funny that we see regular people getting the concept of the cloud while technical folk, who have been using the term for decades, are trying to turn it into something new.

Re:Balderdash

[cforciea]

Bullshit. Cloud computing has had a very specific meaning for a very long time; specifically, it is when you have one virtualized, logical machine that is is running on some arbitrary number of physical hosts. Just because a bunch of asstards have come in and turned it into some almost meaningless buzzword that means vaguely Software as a Service doesn't mean it's always been that way. You assholes should have picked a term that wasn't already in use if you didn't want there to be confusion.

Re:Balderdash

[mini me]

The Cloud is not the same as Cloud Computing. The Cloud is this (Special note: Image is dated 1998). Somewhere in the fog is your service. You, the end user, don't care about how it works, it just does – always.

Re:Balderdash

[WaffleMonster]

The Cloud is an abstract interface, not a specific technology. It always has been. Look at some networking documents from years ago and you'll find the cloud present. The cloud services might be implemented using virtualization, but you don't care, because it is just an abstract network that you throw your bits at.

I think it is funny that we see regular people getting the concept of the cloud while technical folk, who have been using the term for decades, are trying to turn it into something new

LOL clouds on network diagrams ususally point to networks or systems outside of your administrative domain.

Is that all "the cloud" is?

Re:Balderdash

[mini me]

Yes, absolutely. But the key point is that, like the internet itself, it has to always be available and just works. Take S3, as an example. Amazon distributes your data across multiple datacenters in geographically distinct locations. Even if a couple of datacenters were blown to smithereens, your data would still show up just like it always has without you ever having known anything has happened. That is why it is shown as an abstract cloud instead of a single wire to a single computer.

Re:Balderdash

[hawguy]

Bullshit. Cloud computing has had a very specific meaning for a very long time; specifically, it is when you have one virtualized, logical machine that is is running on some arbitrary number of physical hosts. Just because a bunch of asstards have come in and turned it into some almost meaningless buzzword that means vaguely Software as a Service doesn't mean it's always been that way. You assholes should have picked a term that wasn't already in use if you didn't want there to be confusion.

You sound like the guy that used to go around correcting people about baud rates "Bullshit! Your modem isn't 28.8baud, it is 3200 baud and 28.8bps! Baud has had a very specific meaning for a very long time and now your asstards can't tell the difference between baud and bps!"

The fight is already lost, the definition of "cloud computing" is so cloudy now (pun intended) that you have to ask for clarification every time someone uses the term.

Re:Balderdash

[CodeBuster]

should have picked a term that wasn't already in use if you didn't want there to be confusion.

Confusion is part and parcel of marketing, it's part of their bag of tricks. After all, confused and ill-informed people are much easier to manipulate without being detected and even if they do find out that they've been lied too, there's always another mark.

Re:Balderdash

[angel'o'sphere]

The cloud is something new!

The cloud has nothing to do with "network"!!
From a programmers point of view a cloud aware application does not need to know anything about memory, cpu and back storage.

Everything is abstracted away and virtualized independendly.

During deployment you only bind a name to a service, you don't even knwo how the service si running, that is up to the cloud provider.

The service can dynamic scale with load (increase cpu, memory etc. assigned to it) or can be capped to not exceed certain costs.

Yes, on the surface it looks like the cloud is mainly utilized by "Web Sites". Bu that is not the case. Big cloud apps are just enterprise level applications doing some business stuff, having nothing to do with Web and are not accessible by mere mortals like you and me.

Re:Balderdash

[Hognoxious]

You know, I was pretty sure it was on a flowchart template I was given when I started work. I still have it, somewhere.

Apparently I was wrong, but I do recall it being used in the way you suggest, vague memories of some EDI stuff I did many years ago.

Re:Balderdash

[lurcher]

The cloud is something new!

The cloud has nothing to do with "network"!!
From a programmers point of view a cloud aware application does not need to know anything about memory, cpu and back storage.

Everything is abstracted away and virtualized independendly.

During deployment you only bind a name to a service, you don't even knwo how the service si running, that is up to the cloud provider.

Hmm, ok, that sounds great, but what does it actually mean. BTW, try and answer using real words, not just hand waving.

Re:Balderdash

[Anonymous Coward]

Its just a repackaging of main frame principles but don't tell anyone it wouldn't be as flashy.

Re:Balderdash

[cforciea]

That's retroactive etymology. Yes, network diagrams have used clouds to represented abstracted away network architecture and/or the Internet for a really long time. But storing things on a/the cloud originally referred to storing it on a cloud computing cluster. The term was just hijacked by a bunch of people in marketing departments and then retconned by apologists into a term originating from unrelated clouds on network diagrams.

Re:Balderdash

[cforciea]

I know the fight is already lost. I just took issue with GGP's implication that somehow people who knew what cloud computing was before the cloud became a buzzword were somehow incapable of grasping the new, crappy, nebulous term and were therefore trying to alter it.

Not to say my blood doesn't boil every time I see that commercial that somehow implicates "The Cloud" as the means by which some lady is editing digital photos, but I do know better than to spend a lot of effort trying to resist the decay of language clarity when nobody else gives a crap.

Re:Balderdash

[Anonymous Coward]

The fight is already lost, the definition of "cloud computing" is so cloudy now (pun intended) that you have to ask for clarification every time someone uses the term.

Or shoot them until they stop. Has that been tried?

Re:Balderdash

[mini me]

The Cloud is useless without the network, but you are right on your point that things just magically work. That is why it is a cloud; you have no absolutely no idea what kind of voodoo magic is going on within to make it all work, it just does, no matter what happens.

Re:Balderdash

[mini me]

The Cloud means that the service will always be available without you having to provide any fail overs of your own. If a couple of datacenters hosting the application/content are destroyed at the exact same time, you will still be able to carry on as if nothing ever happened. If the usage quadruples during that same destruction, the service will still be humming along as if nothing ever happened.

If you have to worry about failure, you are not using a cloud service. If you are thinking about how you need to prevent failure at all cost, you are likely building a cloud service. Anything else is just a regular service.

Re:Balderdash

[mini me]

It sounds like Web 2.0 all over again. Web 2.0 only ever meant computer consumable content served over HTTP (think RSS, ATOM, JSON, XML, etc.) but since AJAX required Web 2.0 content, it got lumped in. Then some designers, excited by the AJAX possibilities, started making websites that weren't meant to look like a magazine so then design got lumped in.

The goal of Cloud Computing is similar to the goals that marketers trumpet The Cloud to achieve, but The Cloud encompasses much more. It is about the entire stack, including the computers (which might utilize Cloud Computing), the network infrastructure, and the software. This is also what the cloud on a network diagram represents. It is a random void that sends and receives bits. If you have to concern yourself with how it works, you are not accessing the Cloud.

Re:Balderdash

[lurcher]

Maybe so, but the part I was questioning was this bit of rainbows and puppies

From a programmers point of view a cloud aware application does not need to know anything about memory, cpu and back storage.

It may not be the case that it matters where those resources are, but that doesn't mean the programmer doesn't have to know about them.

I don't know the absolute address that the MMU maps a running process under Linux to, but I still need to know about the virtual address that I see in the process space.

Re:Balderdash

[cforciea]

The goal of Cloud Computing is similar to the goals that marketers trumpet The Cloud to achieve, but The Cloud encompasses much more.

Not really at all. Cloud computing is about putting a layer of abstraction between a logical server and its hardware and then running distributed hardware as your physical host. Its design goal has nothing to do with syncing content between clients, or web hosted applications, or whatever else "The Cloud" is supposed to be doing for us. It's (very specific) goals are to provide easy scalability of a logical server and to remove hardware and localized network outages as points of failure.

It is about the entire stack, including the computers (which might utilize Cloud Computing), the network infrastructure, and the software. This is also what the cloud on a network diagram represents. It is a random void that sends and receives bits. If you have to concern yourself with how it works, you are not accessing the Cloud.

That's also crap. The cloud on a network diagram only has to do with software insofar as it relates to network connectivity. The only thing that little piece of clipart relates to is actual data connectivity. The cloud in that picture can't store data, it doesn't manage syncing, and really, it does absolutely nothing on the application layer. It provides layer 3 connectivity from each device connected to all of the others. This is again completely and utterly different from the horrible mishmash term that has arisen in the past couple of years, and the only source for an etymological link is people trying to revise history.

Re:Balderdash

[WaffleMonster]

Yes, absolutely. But the key point is that, like the internet itself, it has to always be available and just works. Take S3, as an example. Amazon distributes your data across multiple datacenters in geographically distinct locations. Even if a couple of datacenters were blown to smithereens, your data would still show up just like it always has without you ever having known anything has happened. That is why it is shown as an abstract cloud instead of a single wire to a single computer

I'm not familiar with the reliability requirement in order to draw a "cloud" on a network diagram. I've seen and drawn hundreds of them myself. Resilliance in the face of disaster was simply never part of the equation. It was always about administrative boundaries.

When did the definition of "cloud" change to imply a reliability requirement?

Re:Balderdash

[cforciea]

Er, you do know that there have been some very public geographic outages of specifically Amazon's cloud services, right?

Re:Balderdash

[AK Marc]

In the 1990s, they called "the cloud" "thin client." Then, in the 2000s, they called it "managed hosting." Now they call it "the cloud." (it was mainframe or terminal computing before that and has had other terms at other times) It's selling failed services to people who wouldn't want them if they used accurate terminology, so they make up new terms. And all "cloud" services are the same and quite confused by those selling and buying them, so it's hard to sit here and make a distinction between "cloud computing" "cloud storage" "cloud services" and "The Cloud" when no one else does. The term has been hijacked and will never mean what it did 10 or 20 years ago.

Re:Balderdash

[mini me]

loud computing is about putting a layer of abstraction between a logical server and its hardware

Which is also true of The Cloud. You have one point of access, but the content/processing is spread across distinct regions of the world in multiple datacenters. Outages at any given location, or even multiple outages, have no effect on service as far as the end user is concerned.

The cloud on a network diagram

In a network diagram the cloud represents an abstract entity that sends and receives bits, nothing more. The Cloud is the same, you send your request, and you get a response back. Going back to my first point, it is represented as a cloud instead of distinct wires because its ability to never fail because of localized outages. It always works, no matter how bad a datacenter is failing at a given time.

Re:Balderdash

[cforciea]

Which is also true of The Cloud. You have one point of access, but the content/processing is spread across distinct regions of the world in multiple datacenters. Outages at any given location, or even multiple outages, have no effect on service as far as the end user is concerned.

Okay cool. So we're in agreement that "The Cloud" is just a hijack of cloud computing terminology turned into a nebulous buzzword.

In a network diagram the cloud represents an abstract entity that sends and receives bits, nothing more. The Cloud is the same, you send your request, and you get a response back. Going back to my first point, it is represented as a cloud instead of distinct wires because its ability to never fail because of localized outages. It always works, no matter how bad a datacenter is failing at a given time.

Damn. Nevermind, I guess.

And you couldn't be more wrong about what that cloud on a network diagram means. If there are no explicit notes on a cloud in a network diagram, it just means that it is abstracted away network infrastructure, and is usually used to either indicate connection to the internet at large or at least to a network that you know will transport your data without having control or knowledge of the specific network topology. There is no implication of any redundancy. There is no relation to distributed computing. It just represents wires that you can't or won't draw because they don't add useful information to your network diagram.

The bottom line is that you can spew rainbows and lollipops at me all day (which business's marketing department did you say you work in again?), but that won't change the fact that a bunch of people in some marketing department someplace stole some perfectly good terminology and ruined it.

Re:Balderdash

[mini me]

There is no implication of any redundancy.

Then why not draw a straight line? The reason we draw a cloud is because we know it is a vast and complex system that allows for faults. The internet itself is designed to survive catastrophe, thus it gets a cloud symbol. If you have one physical line between your two endpoints, even if it is controlled by someone else, there is no need for a cloud drawing in your diagram.

Re:Balderdash

[cforciea]

It isn't a straight line because there might be network infrastructure there beyond a single wire. In many cases, I could probably even draw it all out, but because I want my graph to be informational, I leave it out so that only the the data I wish to convey is described in any detail. It's a way of saying that there is likely additional complexity here, but for the purposes of this graph, you can treat it as a hub and spoke network with a single hub. There is no uptime guarantee or any notion of actually storing data there, unless otherwise enumerated.

Re:Balderdash

[angel'o'sphere]

Perhaps you should become a programmer? Or what ever ;D
If that is for you habdwaving ... you seem to have no clue about computers, OSes and system architectures. What real world word do you need?
Cloud Provider: Google AppEngein, Amazon EC or: the data center in the company I work right now ...
Memory: the little black bugs inside of your computer, usualy soldered on top of a small plastic stribe with golden legs (try to pull out a few and you see your computer stops working ...)
Memory the second, back storage: hard drives, I guess you have a remote idea what that is, or?
CPU: the big thing in the middle of your computer with the huge cooling and fan on top of it, usually soldered on the board, don't try to toy with it.
Virtualization: having something run on a computer, the virtualized thing does not really exist, some ersatz hardware is used to simulate it. E.g. virtual memory is simulated by hard drive memory.
If you need more "words" I suggest to try wikipedia on "cloud computing"

Re:Balderdash

[Anonymous Coward]

We got it when they called it a Network. Now it's a "Cloud"? Stop trying to tell me it's different than a network + software.

Re:Balderdash

[Slashdot Parent]

Bullshit. Cloud computing has had a very specific meaning for a very long time; specifically, it is when you have one virtualized, logical machine that is is running on some arbitrary number of physical hosts.

Methinks you have confused cloud computing with grid computing.

Re:Balderdash

[cforciea]

I have not.

Re:Balderdash

[angel'o'sphere]

And which run mad /. er mods me here as troll again? Rofl ....

Re:Balderdash

[bobs666]

I will not mod you as a troll,

But I will point out that virtual memory is not on the hard disk. Thats swap space. Virtual Memory is an address mapping that lets all the simultaneous processes think they all have access to the same address space.

I think of it as a base address register that offsets a memory page from a process address into a physical memory address. It also has a nice side that memory pages can be moved all around with only the simple change of the base address register.

I am not a hardware guy, this is just what I remember from school, 30 or so years ago. effect that

Re:Balderdash

[angel'o'sphere]

Well, then you should go back to school, in your terms swap space and virtual memory is the same.
And your explanation for virtual memory is wrong, that would be considered "shared" memory ...

http://en.wikipedia.org/wiki/Virtual_memory

Good thing the cloud got delayed today

[afidel]

Well then it's a good thing that VMWare just put up a big old toll plaza on the road to the cloud that will slow things down significantly for many organizations. For those that aren't aware VMWare just announced their pricing model for ESX 5 and it's pretty freaking outrageous, $90/GB list bought in $2800 increments if you want Enterprise with most of the good features.

Re:Good thing the cloud got delayed today

[TooMuchToDo]

Holy shit. Why wouldn't you just go with Openstack, KVM, or Xen at those prices?

Re:Good thing the cloud got delayed today

[jfalcon]

Well then it's a good thing that VMWare just put up a big old toll plaza on the road to the cloud that will slow things down significantly for many organizations. For those that aren't aware VMWare just announced their pricing model for ESX 5 and it's pretty freaking outrageous, $90/GB list bought in $2800 increments if you want Enterprise with most of the good features.

Doesn't affect the big players in the market already who built on XEN which is open and mostly free.

Re:Good thing the cloud got delayed today

[afidel]

You mean the ones that keep having multiday outages? Yeah like I believe that will be 40-50% of new IT spend in only 4 years...
My org is fairly pro SaaS but we just got hit by a half day outage with Salesforce on the golive day for a new solution that of course uses Salesforce as the user interface. That was egg on our face that we couldn't fix and which made quite a few people sit up and take notice. I'm not sure that it's a death nail for new projects being based around cloud/SaaS here but significantly more downtime and it's going to be a serious risk flagged by project sponsors and auditors.

Re:Good thing the cloud got delayed today

[phoebus1553]

I'm curious to see what their SnS cost is now. If you've ever seen year 2 of licensing you know that it's nowhere near the purchase cost for that year of support. Honestly though, they had to do something even though this is waaaaay off the deep end.

Compared to capacity 3 years ago, we could have dropped 30%CPU lics around this year after a major blade hardware refresh from many mixed dual socket 2-4 core, 16-48GB blades to a few dual socket 6 core 96GB blades. Increased capacity and cpu on new sockets is ridiculous and a lot of places probably dropped their support contracts by insane margins. VMWare was going to take it in the shorts due to reduced license needs (CPU) this way they'll get eaten due to walking customers. Maybe they could have found a middle ground... maybe the discounts start off at anything over the old 2CPU threshold and go up steeply after that, after all nobody pays list price ;)

We didn't drop our support at all simply because if we had to re-buy a blade or two of license for any random huge project (which were on the verge of getting approved, and all of them did) it would have stolen all the benefit of dropping those lics for a year. That decision would look to be paying off in spades now since our memory footprint is going to expand HUGE but we're still over-licensed.

Re:Good thing the cloud got delayed today

[afidel]

I'm already at 72GB/CPU physical which means to actually use my existing hardware I'd be paying 225% of my current licensing, and my boxes are tame compared to some configurations like R910 with only dual CPU's but lots of ram or IBM x5 with the memory expander or Cisco UCS with their big memory configuration. My favorite slide was new bigger guests supported (1TB of ram!), with my snide comment of "if you can afford the $112k in licensing just for that VM". I'd say the best compromise probably would have been 128GB per Enterprise license which is equivilant to what you get with Enterprise in ESX 4 with dual CPU's and the 256GB licensing limit. They'd still sell additional licenses as people buy bigger hosts and virtualize bigger workloads. That's actually what I'm going to recommend when I talk to my rep, because I know VMWare is getting an earful today over this BS.

Re:Good thing the cloud got delayed today

[Anonymous Coward]

It is death knell, not death nail http://www.thefreedictionary.com/death+knell

Re:Good thing the cloud got delayed today

[afidel]

It's a contraction of death knell and nail in the coffin that I picked up from some short story I read years ago.

Re:Good thing the cloud got delayed today

[drsmithy]

Holy shit. Why wouldn't you just go with Openstack, KVM, or Xen at those prices?

Because outside of trivial environments it's not a significant expense, and because VMware is the best.

Re:Good thing the cloud got delayed today

[Anonymous Coward]

When did they announce that?

Re:Good thing the cloud got delayed today

[Sprouticus]

Exactly. all it takes is one bad event to make people retract their support.

I am not against SaaS ('cloud' annoys me) for some services. Things like EDI and anti-spam services fit very well in such a model . Data is not stored only transferred, and there is a very narrow scope which makes management easy to handle. It also reduces the scope if an outage occurs.

But when people keep saying that everything is going to be cloud based in 10 years, they just are ignoring the reality of business.

Re:Good thing the cloud got delayed today

[jimicus]

Don't know about Openstack, but KVM and Xen both suffer the same problem.

They provide you with a fairly primitive - albeit effective - toolkit. They don't provide you with a pre-cooked setup which you can just hit "Install" on and 15 minutes later, away you go. If you want to do anything flashy (for instance, put together something that competes with AWS), you are going to have to dedicate insane amounts of time to it.

If you just need virtualisation on a couple of cheapie Dell servers they're fantastic. But if that's all you need, you're hardly VMWare's target market anyway.

Re:Good thing the cloud got delayed today

[afidel]

We're what I'd call non-trivial (240VM's today, up from a handful 18 months ago) and yes a $100k licensing upgrade cost to utilize the exact same resources we have today is VERY significant, plus it increases ongoing costs because you have to pay subscription and support on all those additional licenses. To put that in perspective it's more than 1/4th the cost of our entire farm including hardware with 5 years 24x7 6 hour support, Microsoft licensing, and existing VMWare licensing. All for what amounts to a fairly minor bump in features, that's not a good ROI no matter how you look at it.

Re:Good thing the cloud got delayed today

[Anonymous Coward]

http://www.vmware.com/files/pdf/vsphere_pricing.pdf

Re:Good thing the cloud got delayed today

[DragonWriter]

Don't know about Openstack, but KVM and Xen both suffer the same problem.

They provide you with a fairly primitive - albeit effective - toolkit. They don't provide you with a pre-cooked setup which you can just hit "Install" on and 15 minutes later, away you go. If you want to do anything flashy (for instance, put together something that competes with AWS), you are going to have to dedicate insane amounts of time to it.

I think that's true (by design, really) for KVM and Xen, but Openstack and Eucalyptus are, as I understand it, more complete packages (Eucalyptus is API-compatible with AWS, and OpenStack is, from my understanding, similar in scope to Eucalyptus but uses a different API.)

Re:Good thing the cloud got delayed today

[Synerg1y]

Not really, my co-worker set up Xen in a week on a "dark" server, fully functioning, hosts VM servers (from VMWare rofl), had some nifty console capabilities. We run vmware otherwise with a 6 figure licensing cost that is renewed once in a great while. The problem is nobody knows Xen server and its hard to hire somebody cost effectively that does. I think everybody on here is talking about vastly different environments with different cost analysis scenarios.

Re:Good thing the cloud got delayed today

[randyleepublic]

death "knell", not "nail".

Blame IT for this.

[Anonymous Coward]

The Cloud offers a lot of advantages over conventional IT:

o Scalability when you want it.
o No need to deal with the payroll of admins.
o Guaranteed service levels.
o Actual security. There has yet to be a breach with a cloud provider. Plus, this is what SLAs are for.
o Money savings by not having to have a server room, with the attendant HVAC, power, and security requirements.
o Not handcuffed to OS, application, and database versions.
o No need to concern about if running the latest Exchange versions.
o No need to deal with backups and the attendant infrastructure.

For a midsized company, one would be a fool not to go with cloud based services. Cloud security has shown itself to be excellent. The only people who complain about the cloud are those with a vested interest in the old style of running things. Kick the old server guys to the curb -- technology moves on, and businesses don't need to have server rooms anymore, just like businesses don't need TTYs and printers attached to terminals. Businesses don't even need to care what a server looks like. The cloud provider which has advantages of economies of scale can deal with that.

Re:Blame IT for this.

[tokul]

o No need to deal with the payroll of admins. o No need to deal with backups and the attendant infrastructure.

Oh really. How do you connect to that cloud thingy. You still need some local IT infrastructure for that.

o Guaranteed service levels.

Contractually guaranteed. Good luck with that contract, when your business depends on third party, third party fscks up and only you have is some paper.

o Actual security. There has yet to be a breach with a cloud provider. Plus, this is what SLAs are for.

http://en.wikipedia.org/wiki/Dropbox_(service)#Criticism

o Not handcuffed to OS, application, and database versions.

Only if your cloud service provider does not have some dependency on client OS.

Re:Blame IT for this.

[WaffleMonster]

For a midsized company, one would be a fool not to go with cloud based services. Cloud security has shown itself to be excellent.

Hosting providers are hacked all the time.

The only people who complain about the cloud are those with a vested interest in the old style of running things

The only people who complain about eating worms are those with vested interests in agriculture.

Kick the old server guys to the curb -- technology moves on, and businesses don't need to have server rooms anymore, just like businesses don't need TTYs and printers attached to terminals

I like economies of scale. I dislike blanket statements made from a position of extreme ignorance. You have no clue what all "midsized companies" requirements are -- keep your judgements to yourself.

Re:Blame IT for this.

[Anonymous Coward]

I love the idea that the "cloud", as you so cheerfully put it, has no servers. It is a magical land where all computer problems turn into puppies and rainbows, for FREE! So either this magical land pays for servers, HVAC, redundant power etc (and marks it up for profit) or a company does. Ain't no free lunch, although your little spiel is cute.

Re:Blame IT for this.

[mlts]

I'll bite.

Scalability: Depends. Some things may scale like putting files for download with Akamai. Other things, not so much. The cloud is a tool in the architect's toolbox.

Admin payroll: Someone is going to need to architect things, so you will still have an IT department, if only to maintain your internal network links, and make sure the desktop PCs don't become botnet clients. Don't forget networking. You will need to have those fat pipes to access the cloud provider, so you will still need a space for those Cisco Nexus 7000 boxes with the clear covers.

Guarenteed SLAs: Read the contract... the SLA may not be as good as people think it might be.

Oh, and if the cloud provider goes bankrupt, the data on their servers is free for the taking by anyone who so desires. Even though it is SOP for auctions to wipe the drives, this gets forgotten or "forgotten". Promises mean nothing when some competitor in Elbonia now has your payroll (sold your employees' info to ID thieves), your client list (and is offering the same services for 50% less), your source code (with a version of your product that is exactly the same except in name), your suppliers (which are then harassed to make better deals elsewhere), and others.

No recorded breach with a cloud provider... time will tell on this one. Nothing is 100% secure. Gmail has had people report incidents, Dropbox has had the security tokens that people talked about, and so on.

No server room? I'm sorry, but even with "economies of scale", you are either paying for a data center in house, or you will be paying for one somewhere else. Don't forget regulations about physical security of data.

DB/OS versions? Sure. However, if something breaks your app's code due to an update done without notice, there is no way to roll back. ITIL 101 here.

Exchange versions? See above.

Don't assume the cloud provider has backups automatically. If someone logs on as an admin with a cloud provider and blows all your storage away, it may not be recoverable, while the old LTO-5 tape library will be able to restore data. When push comes to shove, and in some industries, you better retain data for a while (up to 50 years if dealing with the FAA), you need to pack your own parachute. I trust tape, and moving archived data to the latest archival version every couple years far more than a cloud provider's promise. However, I'm one of those "IT guys" that the parent apparently dislike, so if someone wants to be fast and loose with their data, they can store it on the cloud and assume that their storage has all the snapshot features of the EMC SAN they want to chuck.

Cloud computing is a useful tool. It won't replace server rooms anytime soon. Maybe I'm a fossil, but I rather trust a VTL, replicated SAN, or even good old fashioned tape far more than I would trust just an assurance that a cloud provider has my data backed up.

Re:Blame IT for this.

[X.25]

The Cloud offers a lot of advantages over conventional IT: ...
o Actual security. There has yet to be a breach with a cloud provider.

Are you a retard?

Re:Blame IT for this.

[BeerCat]

Bravo! Your post (and the parent) neatly encapsulate both the "for" and "against" arguments. For some businesses, the points you raise will be paramount, and hence they (should, provided the PHBs don't get involved) stick with their own servers and staff.

For others, the lower initial costs will be the attraction, and hence they will go Cloud.

Depends on the needs of the business, though.

Re:Blame IT for this.

[angel'o'sphere]

Dropbox is not a cloud service provider.

It is a company providing the Dropbox service, big difference!!

How they do that, I personally don't know, perhaps they run on Amazon EC or on the Google AppEngine. Google or Amazon would be the cloud service provider in thsi case.

Re:Blame IT for this.

[lurcher]

Actual security. There has yet to be a breach with a cloud provider. Plus, this is what SLAs are for.

Until the US Gov comes along and asks the provider to show it your data (see other slashdot topics).

Re:Blame IT for this.

[Anonymous Coward]

Contractually guaranteed. Good luck with that contract, when your business depends on third party, third party fscks up and only you have is some paper.

There are these things that give contracts teeth. They're called "courts of law." Google it some time.

Re:Blame IT for this.

[Lunix Nutcase]

Contractually guaranteed. Good luck with that contract, when your business depends on third party, third party fscks up and only you have is some paper.

Yeah, it's not like contracts are legally enforceable agreements between parties and you can sue the other side when someone breaches it. Oh wait, they are.

Re:Blame IT for this.

[Lunix Nutcase]

Yeah, you sure showed him with that plethora of examples of Amazon EC2, Rackspace's cloud services, Dell's cloud services, etc being breached. Oh wait, you didn't provide a single one.

Re:Blame IT for this.

[JohnRambo]

1) Cloud Computing is the next "fad" to come down the pike from the marketing departments of companies selling hardware and services. Examples of previous "fads": client server computing, n-tier thin client server computing, web services, virtual servers, etc.... While all these previous technologies have filled a need in the corporation, they also have failed to reach the hype that the marketing machines have set out for them. I predict the same for cloud computing. Some companies will move their applications into a vendor provided cloud, others will build their own internal cloud and keep everything in house. One vision does not fit all. 2) As far as the changes in job roles. If a job becomes easier to perform because of improvements in tools, then those jobs will move down the value chain and will be performed by those with less skills and less pay. So if you want to keep a high paying job, then continuous learning of new technical skills is required for IT. Difficult to do, high pay, easy, lower pay. Dont like that? Change careers. That is the way it has always been in IT and the way it will always be. .

Re:Blame IT for this.

[tokul]

Yeah, it's not like contracts are legally enforceable agreements between parties and you can sue the other side when someone breaches it. Oh wait, they are.

Will court fix problem that you have with service? I don't think so. You can get compensation, but you won't get your customers or reputation back.

Re:Blame IT for this.

[scottbomb]

Or rips away the server in the middle of the night because someone else was hosting kiddie porn.

Re:Blame IT for this.

[AK Marc]

The cost is never lower with a cloud. People only think so when they ignore things like risk. And if the cost of the risk were actually zero, then you could take all your infrastructure out, burn it, and replace it with nothing and save even more money (Aside from the ticket for unauthorized trash burning).

There's a total shock...

[fuzzyfuzzyfungus]

"The Cloud" is really just the latest advance in the relentless encroachment of "appliances", which are just the IT-specific implementation of the replacement of skilled tradesmen with capital-intensive systems and disposable peons that has already done a pretty good job in other industries.

You can replace thousands of jacks-of-many-trades smalltime sysadmins with a few architects and a bunch of screwdriver monkeys. ROI, here we come! (Even the confusion over what constitutes a "cloud" arguably shows the progression in finer detail: Things like EC2 only abstract away the hardware and interconnect stuff, while leaving you with the need for VM admins to actually turn the cloud into services. Things like Azure or Google's App whatever it is abstract away the sysadmins and leave you just needing the coders to write the applications. Hosted applications, webapps, 3rd-party email providers and the like abstract away the apps, and just leave you to point the client at the right URL. As soon as we all get our Chromebooks, we can fire everybody but the licensing person and the janitor, and each replacement laptop will automatically be provisioned according to the spreadsheet maintained by the licensing person as soon as the janitor plunks it on top of the RFID fob built into the desk...)

On the (very bleak) bright side, we might at least get to enjoy a little righteous schadenfreude when the axe comes for those techie-uber-libertarians who have spent years watching other peoples' creeping unemployment with the smug conviction that they are too good for that, and the peons can always retrain for the new jobs that the invisible hand of innovation will shortly be providing...

Re:There's a total shock...

[DigiShaman]

IT technology and its pseudo-toys (a "want" over the "needs" such as the iPad) are still a moving target within the industry. As such, there will always be a layer of complexity and complications that will require a support staff. We can argue whether or not that staff is internal, or provided by an MSP (Managed Service Provider). But none the less, IT support will always be around and thus so too will there be job openings to fulfill that role.

Re:There's a total shock...

[Anonymous Coward]

I thank you for my new favorite word: Schadenfreude.
–noun
satisfaction or pleasure felt at someone else's misfortune.

Who do you trust?

[khasim]

The problem with the "cloud" is that you put your complete trust in
a. the cloud provider
b. the telco that connects you to the cloud

As anyone who has ever had to deal with outside vendors knows, they have no real commitment to your business. You are a single account.

When your business cannot connect to the systems, it is a crisis for your business.

For them, it is another day in the office.

Re:Who do you trust?

[FooAtWFU]

If it comes down to telco downtime, you can always get a second telco. Cable and DSL, or if you can't manage that maybe something cellular/wimax.

Just saying - if that's the only reason for skepticism, you can mitigate it.

Re:Who do you trust?

[mini me]

Absolutely. These days, people at home, let alone businesses, are operating at least two active connections to the internet (cable/DSL and cellular). The internet was designed right from the start to support multiple connections so that a link can fail without anyone noticing. It seems strange to me that network outages are still a topic of debate.

Re:Who do you trust?

[wmbetts]

I was about to say that. I telecommute for my job and it's not a requirement (duel connections), but I do have 2 connections pulled into my house. If for some reason both fail I can tether my cell phone. The back up connection is actually pretty cheap, because it's the basic plan, but would work if I was in a bind.

Re:Who do you trust?

[antifoidulus]

duel connections

No wonder you are having so many issues connecting to the internet.... Your connections are constantly firing pistols at each other at 10 paces.

Re:Who do you trust?

[gullevek]

After the big earthquake in Japan we see this different. Electricity won't go out in a worldwide cloud net, compared to the servers that might be in a location that is affected by possible lights out plans ...

Re:Who do you trust?

[wmbetts]

mod him +5 funny

You're right I misspelled that. Your comment actually made me laugh once I realized my mistake.

Re:Who do you trust?

[yarnosh]

As anyone who has ever had to deal with outside vendors knows, they have no real commitment to your business. You are a single account.

Except when they do. Through an SLA. Also, they have a reputation to maintain. It just isn't good business to go letting your clients flap in the wind.

Now, personally, I absolutely hate relying on outside vendors for support when I know that I have the expertise to handle most problems myself given the right access. But, to someone who doesn't have that expertise, it doesn't really matter a whole lot if the support is outside or inside the company. I don't think it is as bleak as you paint it. You can get good service providers and you can get bad ones. Just as you can hire good local admins or you can hire bad ones. You just have to weigh the costs and risks.

Re:Who do you trust?

[Oceanplexian]

Failover with multiple connections is far from simple, and generally out-of-reach for most businesses. If you're hosting web services, then you need to need to acquire IPv4 space (not exactly easy nowadays), a BGP prefix, an expensive router, and pay hundreds to thousands of dollars a month to >1 ISP that support BGP.

Alternatively, you could just get 2 cheap internet connections and a router that supports active fail-over/load balancing, however now half your address-space on the other ISP is unreachable. Not to mention that those routers cost thousands of dollars if you don't enjoy hours of BSD hacking...

So yeah, it's not that strange that bill-the-office-manager isn't running a HA configuration.

Re:Who do you trust?

[tubs]

And of course, the wires from your premesis go through the same ducting - so that guy with a digger neatly cuts off both your main and backup linls.

Re:Who do you trust?

[sosume]

You can buy dual WAN failover routers for a few hundred bucks .. combine that with two cheap DSL lines, and the cloud is always in reach (provided that the two ISPs use different POPs)

Re:Who do you trust?

[TheRaven64]

It depends on the business. I've had a server hosted with a couple of small colo businesses for a while. In both cases, they employed under a dozen people, and if I had any problems then there was always someone available to fix them. That someone would always be a person who was empowered to make decisions, and whose income depended on keeping customers happy. On the other hand, I don't have the contact details of anyone in a management position at my ISP. If I call them, I talk to someone in an outsourced call centre (run by IBM), who can't make any decisions because they do not represent the company in any capacity.

Re:Who do you trust?

[TheRaven64]

Some years ago, I picked up a dual processor (way before dual core was common) machine on eBay very cheaply, by searching for 'duel processor'. Everyone else was searching for the correct spelling, and so no one else bid on it.

Re:Who do you trust?

[hairyfeet]

And when farmer John cuts the trunk? In my area we were down for nearly a week, we are talking nearly 60,000 people affected, thanks to some farmer digging with his backhoe and tearing up the fiber. Can you afford your business to be down that long?

The only way I would recommend that a company go to the cloud is if they have redundant systems on site, at least enough that the world wouldn't end if someone took out the fiber. Shit happens folks and you'd hate to have your business be screwed simply because you weren't ready for the occasional SNAFU.

Re:Who do you trust?

[jimicus]

Alternatively, you could just get 2 cheap internet connections and a router that supports active fail-over/load balancing, however now half your address-space on the other ISP is unreachable. Not to mention that those routers cost thousands of dollars if you don't enjoy hours of BSD hacking...

pfSense is your friend here.

Re:Who do you trust?

[Anonymous Coward]

Not if you're in a building designed properly.

Re:Who do you trust?

[Maxo-Texas]

You need to be really careful about SLA's.

There are horror stories about SLA's.

For example: No more than 1% downtime per year sounds good-- but that allows them to have you down up to 3 days.

A friends million dollar printer went down. Other bigger customers with SLA's were also having problems at the same time. They got service after about 36 hours. Sure- they got some SLA fine money. Lost a lot more.

Could have ALSO happened if they had an inhouse repair person that got sick/died, etc. But SLA's are just a money deal at the bottom. Sometime it's worth it to the vendor to violate them.

Re:Who do you trust?

[elrous0]

In John's defense, he really needed to dig that ditch. And he don't understand them fancy electrical hoses, what with all their signal pumps and all.

Re:Who do you trust?

[h4rr4r]

Both of them running over the same lines that one car vs telephone-pole or man with a backhoe will take out at once.

Re:Who do you trust?

[Anonymous Coward]

Well.... you already trust b) since they connect the your existing datacenter to rest of the world.
And sometimes a) is more responsive that the local IT dept.
Furthermore, "as anyone who has ever had to deal with" the local IT folks...they have zero commitment to the business and they blame somebody else instead of fixing the problem.

Re:Who do you trust?

[memyselfandeye]

It's amazing that people don't understand this. I work from home and have a dual-wan, only as a fail-over in case something happens on the ISP end. It doesn't do squat when the sewer company knocks off my 'tubes by 'missing the mark', so to speak.

The point of 'The Cloud' isn't to provide redundant on-demand services to a single location. The point is I can take my laptop to Panera and still have access to my files and e-mail when the power goes out in my home/business. Or even better, I can go buy a laptop when mine is lost/stolen/broken and still have access to my files 1,000 miles away without having to worry about a VPN on some dinky 'business DSL' line, or hope that the office IT guy isn't away on vacation the minute we lose power to a server room and my laptop breaks.

The point of 'The Cloud' is that I can spin up servers anywhere in the world, either internally or via a partner/vendor, and quickly have new or redundant access to external and internal services required to do business. Personal PC backups are nice, but some of us don't have time to wait hours for a new hard drive to be installed and cloned.

So yea, I don't trust Rackspace alone. But having a Rackspace and EC2 account is significantly safer and more reliable, and way cheaper, than having a Technet account, a server cabinent, and an internet connection or two (no matter how fast they may be).

Re:Who do you trust?

[Vancorps]

Way cheaper for a single server perhaps, I haven't found it to be any cheaper when you're talking 5-10 servers or more. The cloud is a decent option for the ultra small and the ultra large companies out there. Anyone in between and it's still very much up in the air as to whether it is cheaper or not.

I've found the key is to have enough local infrastructure to function on your own and use cloud services to scale out. So I have a redundant database environment locally which replicates to the cloud when I need additional traffic for our website. Same story for web servers, for any other type of server I haven't found a need to throw it into the cloud.

All the while I fight the urge to call it all bullshit as the added work and security risks involved with integrating a third party into your environment adds to the costs. A few terabytes of space in the cloud will also cost a pretty penny.

Re:Who do you trust?

[dave562]

You're confusing the subject. What you described is hosting. If a company is in the business of hosting applications, they are probably already in some sort of cloud or hybrid-cloud setup.

Your average small business that is moving "into the cloud" is not in the business of hosting their own applications. They only redundancy that they need is another internet feed. Even a cheap-o Sonicwall can handle fail-over.

As much as people whine about how complex and expensive these setups are, they really aren't. For $30,000 a month we get 192 sq/ft of data center space, power, 20MB commit / 100MB burstable feed with FULL redundancy (incoming feeds, firewalls, load balancers and switches). $30,000 is nothing compared to the revenue that we are able to generate, and honestly, we're paying a little bit more than we should be. When the contracts are up for renegotiation, we will be paying even less.

Re:Who do you trust?

[dave562]

The cloud is a decent option for the ultra small and the ultra large companies out there. Anyone in between and it's still very much up in the air as to whether it is cheaper or not.

Anyone doing IT should be doing some sort of "cloud" setup. By that I mean (mostly) virtualized, SAN backed and replicated to a backup site. The question most people have to consider is whether or not it is more affordable to do it yourself, or to lease resources from someone else (a "cloud" provider).

A few terabytes of space in the cloud will also cost a pretty penny.

Storage is what it comes down to for us. I'm running about 75TB worth of storage and bringing hundreds of gigabytes of new data online every week. One of the vendors that we use for Citrix support is transitioning into being a cloud provider. We talked about using them to host our environment for all of five minutes. Once they figured out that our SaaS environment is almost as big as their entire cloud environment, the discussion was over.

Re:Who do you trust?

[AK Marc]

Of which the number is minuscule.

Re:Who do you trust?

[AK Marc]

"The Cloud" is for everyone. Why? Because there is no definition of "The Cloud." Depending on who is selling what, it's SaaS, thin clients, managed hosting, or a host of things everyone has been doing for 10+ years (though maybe not with the level of variety as now).

Anyone doing IT should be doing some sort of "cloud" setup. By that I mean (mostly) virtualized, SAN backed and replicated to a backup site.

I've been doing that for 15+ years for medium sized businesses. The "cloud" is nothing new, but it's just using a confusing term in order to generate a greater push for outsourcing. It's not used so much for description of internal systems because the made up term is generally used by contractors to fabricate reasons to outsource, but because it's a meaningless term, you could move your refrigerator into "the cloud" and you have just virtualized your food distribution.

Re:Who do you trust?

[Vancorps]

Sounds like you're in the same boat as me, I have 100TB of back-end storage. All of that doesn't even necessarily matter as this business is driven by the need for fast access to that content which includes HD video from our shows, having to download all of that to start editing would suck majorly.

I agree with putting "cloud" in quotes as it depends on context to what you mean. I was asked by the CEO about moving services to the cloud, I mentioned to him that our Xen setup with Citrix provisioning services among many other tools means we already have a private cloud so we would not gain a whole lot by moving our stuff. The only time I see it as potentially worth while is when you know you'll have huge traffic spikes and you can preprovision inside Amazon or whatever provider you like. Even then I'm finding Limelight, Akamai, and even Microsoft have better options available from a CDN perspective which scales far better for web services especially if you're doing a cost comparison between Limelight and Amazon EC2. Management and total costs are significantly lower going the Limelight route.

As always, it just depends on your situation, for some people it'll work just fine, for others it won't.

Re:Who do you trust?

[Anonymous Coward]

Or banjos.

Re:Who do you trust?

[Slashdot Parent]

Except when they do. Through an SLA.

SLAs are nice, but they only guarantee that you'll get uptime or a payout. Most of the time, the payout doesn't even come close to compensating the business for its losses. When the stuff hits the fan, you're still just a single account. Bigger accounts will still get the priority. SLAs are nice, in that they provide at least some incentive to the vendor to maintain uptime, but you should definitely assume that the uptime guarantee will be broken every once in a while, and you should plan for it.

This is why you need disaster recovery. SLAs can be broken. Vendors can go bankrupt. Life happens, and you need to plan for it with backup vendors.

Is a new California gold rush...

[ibsteve2u]

Corporations without any IT staff interfacing with "vendors" who are highly proficient at making soothing technical noises while cranking out large bills? Wheeee...money growing on MBA-shaped trees.

And "What trade secrets?" is the least of it....CEOs better start being careful about the content of the emails they send to their mistresses, 'cuz leverage is leverage.

Upon further consideration

[Anonymous Coward]

Private clouds will win in the end. With public clouds you get:

1. No hardware control.
a) You have no control over your server hardware. It could be running on counterfeit bits of string and chewing gum from China for all you know (try explaining that to a defense contractor).
b) You have no control over physical access to your hardware. You'd better hope the guy they hired at minimum wage to watch the door at night didn't get a better offer.
c) You have no control over bandwidth and connectivity agreements. I hope the ISPs between you and the datacenter are friendly, because a single peering dispute with you on the other side or a new and unfriendly QoS arrangement and you're fucked.

2. No cost control.
a) Presently cloud services like Amazon's do not allow you to set usage caps for your servers. A DDoS at the right time could cost you unheard-of amounts of money, your job (hey, you're responsible for this cloud thing, right?) or even bankrupt your enterprise.
b) Usage-based costs are unpredictable and will play hell with your budget forecasts.
c) Fees for everything. TANSTAAFL.

3. No data control.
a) You have no guarantee your data is securely disposed of when your cloud provider recycles its drives.
b) Your data could go missing at any time for no reason in a manner that prevents its restoration from anything but expensive offsite storage, which you're still maintaining now that you're using the cloud, right? -- As Amazon so eloquently proved.
c) What happens when your cloud provider goes out of business like Iron Mountain?

Private clouds will afford you most of the benefits with none of these problems. In the end it all boils down to money: the cloud has brought us commodity-scale computing, and you get what you pay for. If anything, I foresee a bright future in fabrics, and not the textile kind. Better brush off those networking credentials.

Re:Upon further consideration

[CBravo]

I can't believe you haven't been modded insightful yet... Head on.

Re:Upon further consideration

[Anonymous Coward]

If it weren't an AC post, I would have modded it up. If people want their posts to get more attention, they should log in.

(Posting anonymously so as not to undo mod points already spent. This is just about the only reason I ever post anonymously.)

Re:Upon further consideration

[Ksevio]

I somehow doubt that a small business can provide a better hardware solution with proper distribution and back-up solutions while keeping the cost lower than a professional cloud service.

One would be a fool not to go with cloud services?

[D4C5CE]

For a midsized company, one would be a fool not to go with cloud based services.

Until that close encounter between an excavator and both of your redundant Internet connections... ;-)

hello....

[Anonymous Coward]

Salam kenal

http://asifakarya.web.id

This is silly.

[linest]

The writer believes that "The Cloud" has little/no reliance on networking or storage. That's silly. Infrastructure matters. Pretending it doesn't is perilous. "The Cloud" doesn't change that, it just moves that infrastructure further away from the end user. That doesn't mean it disappears.

Re:This is silly.

[mini me]

The cloud exists to hide infrastructure. The cloud is an abstract network. If, for example, all the service's ties to the USA die, you are automatically routed to the datacenter in Germany. You don't know or care about how it works, it just does. It doesn't fail because it is the cloud.

If you service cannot automatically deal with failure, it is not the cloud, it is just a regular node on the network.

Re:This is silly.

[WaffleMonster]

The cloud exists to hide infrastructure. The cloud is an abstract network. If, for example, all the service's ties to the USA die, you are automatically routed to the datacenter in Germany. You don't know or care about how it works, it just does. It doesn't fail because it is the cloud.

If you service cannot automatically deal with failure, it is not the cloud, it is just a regular node on the network

The core issue with this entire conversation is that nobody has any idea what "the cloud" is. There are only nebulous defintions and marketeers going crazy to apply a new buzzword to their datasheets.

When I hear someone say "the cloud" I immediatly think "idiot" ... The word and usage by itself is too ambiguous to have any concrete real world meaning that makes it in any way a useful term.

At least terms such as "grid computing", "hosting provider" and "Internet" convey some level of understanding.

Re:This is silly.

[mini me]

The cloud has always been this: http://3.bp.blogspot.com/_K52n2Jkv5-I/SYbEngJYXbI/AAAAAAAABVU/C_Wd8gyjaK4/s400/CLOUD.GIF

And that is still what marketers are selling the cloud as, even today. It is an abstract place where things just magically work, nothing more.

VDI [virtual desktop infrastructure]

[zbobet2012]

They say that, in the article. And this, this is why you shouldn't believe a word of it. Its called a Remote Desktop or thin terminal if you are old school enough. Inventing new terms for 30+ year old technologies with links to explain them means this is either a fed PR piece or the author has no idea what he is talking about... or both.

yep

[codepunk]

Loaded and migrated well over 300 servers this year from in-house data centers to EC2. You cloud skeptics can keep denying while I keep migrating.

Re:yep

[mini me]

EC2 is not the cloud. EC2 can provide you with the tools necessary to build a cloud service, but by itself it just a datacenter full of computers. And you cannot throw your application on a few EC2 instances in the same datacenter and call it a cloud application either. That is not the cloud, that's just a networked application.

If a service cannot survive simultaneous catastrophic failures in multiple physical locations, it is not a cloud service. Without being intimately familiar with it, I would like to say that Amazon's S3 service would fit the bill for being a cloud service. Given what Amazon has said about it, it does sound like it meets the criteria of the cloud. EC2, however, does not.

Re:yep

[afidel]

Actually S3 has the same failure domains as EC2 unless you pay $$$$$ for multizone replication.

Re:yep

[Anonymous Coward]

Nope.

Amazon S3 is nothing but storage. This is the same mistake made by people who believe Dropbox is "the cloud." Dropbox and S3 are no more a "cloud" application than it would be if I setup a RAID0 between 2 drives, each with 500GB of space, in 1 server and stuck them in my basement, then gave you an account with 800GB of space that appeared to you as one drive.

Amazon EC2 is the true cloud application. EC2 is a remote-accessible VM. Essentially, it's a reusable application, the specific user data of which is stored off site, along with the application.

Think of this in terms of Google Docs. If Google Docs just gave you a place where you could upload PDF files, and you could allow others to download them, and that was all, that wouldn't be enough to reach the minimum bar to be "the cloud." However, having the ability to edit those documents, in any OS, in (almost) any browser, from any location, because the same, reusable software and your specific user data are stored together, off-site, that's the cloud.

No matter how complex or feature-packed, storage space is NOT the cloud. This is a common mistake people make all the time. The cloud requires more than that.

Re:yep

[Anonymous Coward]

EC2 is not the cloud

Err, sure it is. "the cloud" is a stupid buzzword. It means "anything that is not on local hardware". Email is "cloud computing". You think I'm wrong ? The thing is, the word has no other meaning anymore if you read normal news, and it should be treated as such. What you want to redefine "the cloud" to be, you should find another word for it. Like "secure cloud" or whatever. It will of course only work for a while until that word is watered out to mean nothing too, but that's another issue.

Re:yep

[Anonymous Coward]

Loaded and migrated well over 300 servers this year from in-house data centers to EC2. You cloud skeptics can keep denying while I keep migrating.

...and then what, after you finish that and they're done with you? You probably don't think that will happen, but it will, and right now you're contributing to more decline. You really hate your career and others who make a living in IT that much that you're happy helping to end peoples' jobs?

Re:yep

[Anonymous Coward]

The definition of cloud is dependent upon context...and using NIST's definition...Amazon's service offernings (EC2, S3, etc.) do reflect a cloud environment. The key to building reliable solutions is very heavily dependent upon the Architecture. Amazon's EC2 and S3 services don't exist separately, and share much of the same data center infrastructure, that allow applications to be built to satisfy fault tolerance and high availablity requirements. You can in fact implement resources across multiple zones for relatively little money compared to what it would cost you to build your own HA/fault tolerant solution otherwise. Additionally, you can snapshot running EC2 machines, monitor a running system, and have EC2 servers automatically restarted if there is an event...in another availability zone (which is a separate data center). As a counterpoint...I don't think you'll find any data center environment that can guarantee that their services will never go offline (regardless of whether it's managed by your company or outsourced). Thus, I'd feel pretty warm and fuzzy being able to take advantage of a cloud provider that can offer multiple data centers that are grouped in a pseudo-mesh like configuration so that I can deploy my servers/applications in a manner where they are accessibile even if they do have an availability zone go offline...and the EC2 service can be deployed and managed in this manner. Once again...your degree of assurance around reliability, integrity, and the availability of your business applications will be dependent upon the architecture.

Re:yep

[Anonymous Coward]

EC2 is "Infrastructure as a service" which is a cloud service. What you define as cloud is 'platform as a service' but it is just another cloud service.

Re:yep

[Anonymous Coward]

The NIST Definition of Cloud Computing
        * On-demand self-service; Broad network access; Resource pooling; Rapid elasticity; Measured Service

Yes, EC2 is cloud computing. For that matter even by your definition "survive simultaneous catastrophic failures in multiple physical locations", EC2 is still cloud computing: multiple datacenters can fail and *EC2* can continue to survive. The junk that customers build on top of EC2 does not magically meet either definition of cloud computing, but EC2 (the Amazon service) meets both.

Re:yep

[DragonWriter]

EC2 is not the cloud.

  "The Cloud" is a vague term which is sometimes used to refer to cloud computing (whether or not remote hosting is involved), sometimes used to refer to simple remote application hosting whether or not clould computing is involved, and sometimes used to refer to web-based applications, whether or not remote hosting (from the business to whom the app belongs) or cloud computing are involved.

EC2 is an Infrastructure-as-a-Service (IaaS) offering using cloud computing (dynamic on-demand logical server provisioning, which abstracts, from the user, the underlying physical servers) technology.

If a service cannot survive simultaneous catastrophic failures in multiple physical locations, it is not a cloud service.

That's certainly a kind of service that is facilitated by cloud computing technologies, but services like that existed before cloud computing, and not all applications of cloud computing are intended to provide that particular feature.

It's about the points of failure.

[khasim]

When the systems were in your office, you had X number of points of failure.

And you had someone on site who's job it was to make sure that those systems were available to you.

So you're moving to the "cloud" to save money ... by increasing the number of the points of failure.

So then you add additional systems to mitigate the problems that come with the "cloud". And you probably outsource the maintenance of those systems as well.

And everything is fine until there is a problem. At which point you realize that all the people who you depend upon to keep your systems working only see you as another account. If your business fails, then they're out a portion of their income (until they replace the account with another one).

Re:It's about the points of failure.

[Oceanplexian]

When the systems were in your office, you had X number of points of failure. And you had someone on site who's job it was to make sure that those systems were available to you. So you're moving to the "cloud" to save money ... by increasing the number of the points of failure.

I think it's a fallacy to assume that you're increasing the points of failure by moving to the cloud. Most cloud-based services are hosted in a collocation facility where the provider has controlled humidity, temperature, and multiple redundant uplinks.

Meanwhile, the average company running a mail server in a closet has tons of potential points of failure, not to mention the upkeep of the hardware.

Sure, I'd like to think that "we" can do it better than the big guys, but for 99% of the small-not-tech crowd, it opens a lot of doors and improves reliability.

Re:It's about the points of failure.

[TheRaven64]

Cloud is a bit of a buzzword that isn't properly defined. The idea of a 'private cloud' is quite interesting, because it allows you to separate IT responsibilities. Some people are responsible for ensuring that there is capacity available to do stuff (but don't have to know nor care what that stuff is). Some people are responsible for making sure that the stuff that your business cares about works. If a department needs a mail server, then they can just chuck a mail server virtual appliance onto your infrastructure. Load increases, and the people managing the back end need to make sure that they can cope with this.

This is very similar to the old VM/370 stuff from IBM. You replaced a minicomputer in every department with a mainframe, and if a department needed more server capacity then they just got some new VMs spawned. The difference is that you're now using relatively cheap commodity hardware.

If you care about availability, then you use something like the high availability mode in Xen 4 to keep your VMs in sync between two data centres (which can just be racks at opposite ends of the building connected to different ISPs - or just connected to your LAN at different points). If one machine in either rack fails, then the VMs in it fail over to the other one and are then live migrated back to the other machines in the first rack and you're back to running a redundant copy, you've just lost a bit of your spare capacity. No one outside of the team responsible for maintaining the back-end infrastructure needs to know anything happened - they just keep working.

You can, of course, outsource the back end and the front end stuff. Whether this makes sense depends on the company. For example, if you're a small business in a building full of other small businesses, then it may make sense for you to share the back end provisioning stuff, just as it may make sense for you to have a shared photocopier and receptionist.

Re:It's about the points of failure.

[Princeofcups]

If your business fails, then they're out a portion of their income (until they replace the account with another one).

Ah, but you are forgetting the crucial point. It's not your (the middle manager's) fault that it failed, it's theirs. You discharged your responsibility with the outsourcing, er, cloud contract.

Re:It's about the points of failure.

[Ksevio]

If it's actually a cloud service (not just a server) that you're getting, then it should be distributed and backed up properly, thus reducing the single points of failure. Furthermore, it will be around multiple geographical locations, so even if a tornado comes by and takes out your physical office, the servers will still be running.

The cloud computers will also be in a professional data center with people dedicated just to keep it running. Chances are the IT guy in the office has other responsibilities, even if he is a pro, and the office data center is much less likely to have fallback systems than the full data center.

So, everything is fine, even if your office loses power and a squirrel chews through the fiber.

Re:It's about the points of failure.

[AK Marc]

I think it's a fallacy to assume that you're increasing the points of failure by moving to the cloud. Most cloud-based services are hosted in a collocation facility where the provider has controlled humidity, temperature, and multiple redundant uplinks.

People are being pressed to move to "the cloud" for internal-only applications. It doesn't matter if the managed servers holding their data are all up, if their office is down, then they are down. But the people selling "the cloud" never mention this fact. So the main office gets stripped of redundancy and power backup, and you are left with greater cost and lower reliability than on-site servers.

The cloud is more of the same

[Anonymous Coward]

I have been working with the Amazon cloud now for over three years. Same work as off the cloud and you do not get to crawl in the dust and change disks once in a while. Having no hardware is makes life much nicer.
As to reliability, it is up to you and your companies budget to do it the right way; properly using the resources the cloud provides can make your system very reliable.

relying on public cloud is a bad business model

[bgalbrecht]

The thing people forget about the public cloud is that it's metered usage. For the most part, if a business wants to go to cloud computing, they're going to want to have an internal private cloud so they can have controlled fixed costs, and have overflow capacity available in a public cloud for when management decides the overflow is necessary. If they rely solely on public clouds, they're going to have to have careful monitoring of the public cloud's usage. For most companies, heads will roll if the company incurs multiple times normal monthly cloud computing costs because of things out of the company's control, like cyber attacks, stealth use of the company's cloud resources, etc. For a small company, surprise public cloud bills could be fatal.

It may be that a lot of companies will turn to externally managed private clouds, which is not really all that much different from today's managed hosts at remote ISPs. In that case some of the "loser jobs" will probably be migrating from individual companies to the cloud providers.

Heads in the clouds

[Anonymous Coward]

I can see the merits in having a cloud based back up service continually preserving data that I produce during work.

I'd be wary however about migrating my applications to the cloud because I'd then be reliant on the bandwidth of the connection to the cloud which is subject to a range of things outside my control. From this point of view, I can't imagine my company ditching our IT infrastructure any time soon so reckon our beloved IT team are safe for the time being... although I'm sure many of the senior management who are complete technophobes and seem to view technical people with contempt would love to move everything to the cloud so that they didn't have to endure the humbling experience of being helped by a taciturn technician wearing a Slayer T-Shirt covered in doritto crumbs.

Cloud Service TCO is not all it's cracked up to be

[VTEX]

I gotta say, I've tried several "Cloud" services and I am not impressed when it comes to TCO. While it's nice to have on demand provisioning, the performance of the virtual server instances are generally very poor compared to basic dedicated servers that are similarly priced.

Don't believe me.... run a simple sysbench and test the cpu and file io on EC2 (or your favorite cloud service) and compare it to a hosted dedicated box. In general on a similarly priced and spec-ed VM/machine you are going to find orders of magnitude better performance on the dedicated box, and that performance equals money saved.

In addition, many dedicated hosts now have 2 or 4 hour provisioning times - so if you don't need more boxes instantly, it's not as if you need to wait that long.

So yes, in the end, cloud services can be convenient.... but you certainly pay for that convenience. They are great for testing and development, but long-term hosting.... eh.... most companies would be better off if they stopped drinking the kool-aid and looked at other options.

It's not just a single cost

[alphatel]

We have had this discussion with our clients, and none of them has moved to the cloud. A new business might very well be able to start with cloud services and possibly migrate some data inhouse as they grow, but most established businesses find the model prohibitive because:

• Failover to another high-speed line with SLA must be established for continuous connectivity
• The cost of data migration for existing data
• Locating, implementing and monitoring reliable cloud backup solutions
• Security, encryption and password enforcement at much higher levels than in-house
• No single responsible agent for outages, changes, troubleshooting
• Reverting back to unnamed blame game which was eliminated many years ago by any competent IT team

Paying all the above costs, and if the service(s) do not work as expected, paying for a full year anyway, plus the re-migration back to home base.

Typical business view of hands-on work

[dkleinsc]

If we all outsource all the hands-on work, that means nobody will have to do it anymore!

Seriously, this seems to be the way some business folks think. Outsource your call centers to India, and poof, no more call centers. Outsource your development to Russia, and poof, no more developers. Because it's not taken care of internally, it must be that the work doesn't happen or isn't necessary anymore.

cellular? so you are willing to pay $$$ data bill

[Joe_Dragon]

Comcast Business Class NO CAP vs 4g that can have data bills that hit $10 per GB or some that slow you down after hitting as low as 30GB.

"Cloud" not a new concept

[Anonymous Coward]

So has anyone bothered to point out that Cloud technology is not a new idea? We moved away from the thin client to mainframe model starting in the 1970s. Why go back?

Good opportunity

[Anonymous Coward]

When things go tits up because of issues with the cloud, people will move back to running their own servers. We've already billed the shit out of a couple of clients that dropped us to ascend into the cloud. When they began having troubles they came to us to buy entirely new infrastructure.

Basically, we see a lot of people with aging infrastructure that runs like shit and they think ascending into the cloud will solve everything. They don't listen when you explain the cloud's problems. When they discover the problems themselves, they have no other choice but to come back and spend a boatload on hardware and labor.

This is an opportunity for every business-oriented tech company. It's laughable watching these hipster doofaces squirm in the cloud.

Cost and Time

[DarthVain]

Managers are sick of hearing it will take 5 years to develop and 5 million dollars.

They want it done for next Thursday, for 30,000 dollars. If that entails taking on a bit of risk associated with that, I think most would say "OK cool."...

Many IT shops are bloated and inefficient, and outsource everything really technical to consultant vendors who rip off the business anyway. Take out the middle men, take out the consultant gougers.

Personally for small tasks they seem like an OK solution. However for anything marginally complex, I don't think it is quite there yet. The only thing that Managers might balk at is that they have less control.

Anyway in the end, once things do get complex enough and if cloud starts taking over, the jobs will just move to those companies. Someone has to manage, create, administer this stuff. Which is some respects is a good thing.

Re:Cost and Time

[Anonymous Coward]

Maybe I'm wrong here, but the original needs "those annoyed managers" had, have, and will have, derive from wanting more control over things.

A lack of this control means their job.

The cloud removes this control.

In other words, this is without a doubt, a passing fad...

Effects, not affects.

[skyphyr]

IT jobs don't have feelings.

Yeah, so?

[frank_adrian314159]

Kevin's a day late and a dollar short. He thinks that IT can still control "cloud-like" infrastructure via the IT org and mandates. It's not going to be that way. If you can only offer 100MB email quotas on your internal servers, the employees in your organization will start using Gmail to route around your damage. Can't provide a decent CRM experience? They will route around you with Salesforce.com. Are your file servers a mess of access permissions so that no one can share a document with someone else across the organization? Hello Google Docs.

The bottom line is to understand you rplace in the connected world. The purpose of IT is to facilitate exchange of information. If your organization looks more like a blockage or network damage, it will be routed around.

Mediocrity of art

[Anonymous Coward]

Programming and networking have reached boundaries that I would put as art, although some are atrocities to be certain. There is elegance in the way process work flows are setup, maintained, and brought to the uneducated of the craft. But yet there seems to be this continually drive for cookie cutter systems that will ultimately see the death of this almost budding art form in IT.

Re:cellular? so you are willing to pay $$$ data bi

[badran]

The 4g will work if you only need to connect to Mail, Chart and svn.

privacy

[Anonymous Coward]

Some businesses won't be able to outsource everything because of various privacy, company, state or university laws/practices. Some have to have complete control and store certain data locally.

Cloud sucks, I know

[Anonymous Coward]

Everyone intelligent knows the cloud is just another marketing tool for companies to milk you for every little dim they can. instead of buying that $2000 one time pay software, now you're stuck paying $39.99 a month for the next 10 years. And if you want to migrate your data? Too bad! Gonna have to stay with us :)