Why the cloud will (entirely) replace in-house applications

Why the cloud will (entirely) replace in-house applications

Summary: Cloud computing is the latest IT "fad" that isn't a fad. How many times do we have to do this before we get it right?


This article is a direct reaction to my colleague Steve Ranger's "Why the cloud will never (entirely) replace in-house applications" article that he posted yesterday. I'm not blaming Steve for his viewpoint, because he shares it with a lot of people.

I can't tell you how many times I hear that "cloud computing is a fad" or "the cloud is inherently not secure", or best of all, "the Cloud is not a sustainable business model". To all of those, I simply say, "Nonsense". I'd like to say something related to bovine excrement, but I'll resist. Cloud computing is not a fad. It is not inherently a security risk. And it certainly is a sustainable business model. Cloud computing is the latest IT "fad" to become the target of the short-sighted commentary and naysaying that I've observed for the past three decades.

If you look at what's happened over those past three decades in IT, you see that I'm spot on with that statement.

Don't believe me?

How about the personal computer, as my first example. That's about 30 or so years old. How did that "fad" turn out? I can remember analysts and newscasters saying that the personal computer fad is a fleeting fancy, an expensive hobby, or a rich man's toy.

I'll name the fads that I can recall for you, but I'll spare you the details of each along the way (and they are not necessarily in chronological order):

  • Personal computers

  • Mobile phones

  • Windows

  • Apple stuff

  • Linux

  • Virtualization

  • Laptops

  • Tablets

  • Blade servers

  • Telecommuting

  • Cloud Computing

Now look back, if you will, and cite examples of those fads that burned out like an old man's pipe. You can't name one, can you? Nope, you can't.

That's because they aren't fads at all. They're real. They fill a need. They're part of our technological evolution.

Cloud computing is no different from the rest of that list.

It fills a need — a need for low-cost, agile, OPEX-charged computing that's available 24x7 anywhere in the world. For some strange reason, analysts and observers see cloud computing as some sort of "tech devil", or some hellspawn that is very different from what has existed for many years under other names.

I find it both disturbing and humorous to imagine these people penning their negative reactions to cloud computing, when most of them haven't a clue what it even is that they're talking about. It's kind of like people who toss around the term "trans fat". They have no idea what trans fat is. They have heard the term and they know it's bad.

Well, in case you're wondering, I know what trans fat is. I know what cis fat is, as well. I can draw them for you. And I can tell you that cloud computing is here to stay, and that all services, even your operating systems, will exist only in the horrible, terrible, unsecure, devil-spawn fad that we now call the Cloud.

It makes me wonder how progress in this world was ever made with so many people standing about saying, "Nope, that'll never work", or "Don't waste your time with that nonsense". It makes me realize that without those hard-headed, negativity-deaf, determined innovators, there are many things that we wouldn't have today. Here's a partial list of those things:

  • Radio

  • Airplanes

  • Cars

  • Trains

  • Telephones

  • An accurate view of the solar system

  • Television

  • Personal computers

  • Cloud computing

Now to some of Mr.Ranger's points from his article:

At the moment cloud still accounts for a relatively small slice of enterprise IT spending — perhaps no more than five or six percent of the total software market, although one prediction sees this climb to 20 percent by the end of the decade.

Yes, it might currently account for a small percentage, but by the end of the decade, it will comprise 90 percent plus. You have to understand that technology adoption isn't linear. When the first personal computers hit the market, uptake was slow, but quickly and exponentially grew to millions of converts. Same for cell phones, TVs, cars, and just about anything else technology-related that you can think of.

That's partly because companies remain cautious about the new technology, but also because they have significant investments in their existing on-premise IT infrastructure, both hardware and software.

True, but how long does it take for on-premise hardware to become obsolete? Three years? Five years? By the end of the decade, none of the hardware in datacenters right now will still be running in datacenters. That's plenty of time for conversion, and it will happen. By 2020, only a small percentage of hardware (<20) will be left on premise as privately-owned. The cloud will be the next utility, the next commodity, if you will, and we'll purchase it much the same as we do electricity, water, or gasoline — in bulk, and as needed.

Indeed, the rapid growth of cloud (and Software as a Service in particular) has made some question whether on-premise applications have a long-term future at all, or whether all applications will eventually be cloud-powered.

See? Steve realizes the folly of thinking that the cloud is a passing fancy or a fad. He is correct that all applications will be cloud-powered.

But it's also possible that on-premise applications still have some use — and some fight — left in them, at least as far as CIOs are concerned.

They do, but very little, and it's running out fast. CIOs will convert to cloud or be cloud-convinced when the winds of change blow their minds that direction. Everyone will soon have a "cloud initiative", or whatever buzz-term someone invents for it. The bottom line is that if you aren't already researching or planning your "cloud initiative", you're behind the curve.

John Gracyalny, VP IT at SafeAmerica Credit Union, said: "I don't consider the internet stable enough for truly critical functions. I'm also reluctant to trust a third-party datacenter's security."

I don't know John, but it sounds like he needs to get on his horse and mosey on over to the General Store and answer the clue telegraph. You trust third parties to prepare your food, to service your car, and to handle your mail, so what is it that makes third-party trust so difficult in IT operations? I can remember when every company wanted to run its own mail server and web server in house. Why? Why do you need to do that? Are you an email provider or a web hosting company? If not, why not allow someone else, who knows what they're doing, do it instead?

I don't build my own cars, washing machines, or vacuum cleaners. I don't even build my own computers anymore. If you need to feel ownership and control of something, plant a tree.

Some CIOs realize that moving to cloud-based services is part of the technology evolution inevitable. Those are the CIOs to keep. The other ones, unfortunately, not so much. If your CIO isn't looking to take your company to the future in technology, what good is he or she? Perhaps there are some openings for CAOs* somewhere.

Cloud computing is a fact. Cloud computing is a good thing. It is the future of IT for business and for personal computing. Choose a direction now, and head toward it for the future, because whether or not you like it, the cloud will (entirely) replace in-house applications. And it will happen sooner, rather than later.

*Chief Anachronism Officers: In charge of keeping things the way they were yesterday.

Topic: Cloud


Kenneth 'Ken' Hess is a full-time Windows and Linux system administrator with 20 years of experience with Mac, Linux, UNIX, and Windows systems in large multi-data center environments.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Unfortunately fad was the wrong word...

    I have worked in the computer industry for 35 years and have seen many changes. I have read the stories that COBOL would be dead, C++ is on the way out and yet they are still here and powering major systems. In fact I am even seeing a high demand for C++ developers.

    I remember years ago that SOA was the next great thing, what has happened to that?

    Currently its Javascript/HTML5 that is being shouted from the roof tops as the only tools developers need. Has anyone tried developing a major project with and them maintaining, its a nightmare.

    Now we have the 'Cloud', a few years back it was called the internet. I mean really the 'Cloud', whatever next. Oh yes, Big Data. Hasn't data warehouses been around for a long time? Ok someone is going to say that is totally different, yes it is, it is a different name.

    The computer industry loves to re-invent it self and come up with new ways to convince companies to hand over their cash.

    The Cloud is just a nightmare waiting to happen as more companies as persuaded to jump on board. It is a single point of failure for a start. Here we go, AWS, Rackspace, Azure are going to shout that they have multiple data centres etc to provide High Availability. Is any body missed the stories about the these systems going down?

    However the single point of failure is the connectivity from the Cloud to your office. If you loose your connection, you loose your systems.

    Now I am not against the Cloud totally it does have its advantages for the likes of sharing information, email, social media etc.

    It is about balance and any company's CIO that puts all his eggs in one basket deserves everything he gets.
    • Preach it brother, preach it...

      People are stuck in a "if you're not with us, you're against us" over this topic. "Cloud", which is just another way of saying "storing my stuff on someone else's property and in their domain of control" is only one option for data that you want to be mined by the cloud provider, Google, Facebook etc. It's not for serious information that needs to remain within the domain of the creator.
      • Sensitive data

        I so agree with your comments. Cloud computing is simply asking someone else to be responsible for you. It may be easier, even cheaper, but there is a price in asking someone else to be responsible for you; it means that you become under their control, their power. If we want to retain power, we need to also retain responsibility. Yes, software is expensive, but I'm willing to do without some other desire in order to have all of my data inhouse. I have no desire to have Google, Facebook, (or the US Government) control my life. So, I choose to take responsibility for it and that means paying, if need be, for the software I need to manage my own affairs. Allowing a "cloud" to manage your affairs means there WILL a cost at some point, and you may not like that cost.
        • Not necessarily cheaper

          Plenty of reports have proven that "the cloud" is not cheaper when all factors are calculated. In fact, most show that it is more expensive ... specially when taking into account the cost of an outage ... which happens a lot more often than the media report.
          • Not necessarily cheaper

            I agree. I have a number of colleagues who are unhappy with the initial cost of good software and love the idea of paying as you go. But, to be honest, I have some heavy duty software that's lasted for years. If I were to divide that initial cost over the years I've had it, it would equal pennies per month to use. Not everyone needs to have the latest version of something. Often, the newer versions have problems anyway and it's better to let the dust shake out before purchasing.
    • SOA?

      You write "I remember years ago that SOA was the next great thing, what has happened to that?" It became the new standard. Every time you shop at Amazon or read the news at Yahoo, you're using an application built according to SOA principles. Welcome to the future.
    • Well written response...

      up until I read "...If you loose your connection, you loose your systems....". It is lose, as in lost, one letter O not two OOs. Loose is something not tight.

      Not that I agree with all you wrote, still it is a good response.
      • Yep, Ken needs a bigger clue....

        I still do not understand why any well thinking company (oops, that may eliminate some) would want to take any risk of depending on "ethernet" connections (in the greater sense, not of the standard) to always be there when THEY want or need them.... It's hard enough to keep the internal networks up & running, but to depend on nebulous "others" just makes no sense at all, especially for business critical data. There is already a growing pool of evidence that accessability is not "continuous" (Unless, I suppose, one were to purchase their own private data link to the offsite location - and even then...)
    • right on

      This is the same tired old argument about distributed versus local sourcing. Your mention of the data pipe is of course the MAJOR weak link. This is just "thin client" in old-speak.

      Just like the rest of "the new MIRACLE [fill in the blank]" that will change the universe, this is just another tool that will be useful for some and a disaster for others.
  • Forecast

    I mentioned to one of my tech account managers recently that I thought small to mid sized infrastructure departments would be virtually nonexistent in 3-5 years and he looked amazed.

    However, as you rightly point out, adoption is alway slow and will accelerate, that leaves 3-5 year hardware refreshes (for those that have not already started) before we are done.

    What you may have left is a network team, but even then i'm not so sure that we won't all have been given our own 4G (or whatever comes next) hotspot access and be done with hardwired.
    • There will always be on-premise IT

      My Dad was in and out of nursing homes and hospitals for months before he died. I had to actively point out things to nurses and doctors, ask them questions, and ensure that at least in critical areas, he was being well attended to. I'm sure if he was living with me, he would be alive today.

      That is the principle of life you are faced with, when you consider outsourcing critical operations to others, instead of taking care of them yourself. No one is going to do a better job than you, because no is as invested in your success, as you. Outsourcing non-critical operations is fine, or even some very important operations, when you simply don't have the means to do so yourself. I just personally don't see how a mid-sized to large business cannot have an on-premises IT department, when IT is vital to its operation.

      So anyone who tells you that he will do a better a better job than you, managing your critical IT resources, when you have the means to do so yourself, is lying to you! No one takes care of your important affairs better than you, when you have the means to do so. This is the way things have always been, and will always be.
      P. Douglas
  • amen to that ...

    I still can't believe there are Cloud deniers out there, but I guess there are some people who still think the world is flat.

    Cloud Pros
    - cost
    - on demand pay structure (don't pay for what you aren't using, or what you will need tomorrow)
    - repeatable security and centralized processes
    - uptime
    - practically limitless resources ... storage, processing, etc ... available now
    - tons of free resources, help, etc

    Cloud Cons
    - relatively new
    - requires a retooled skill set

    AWS, Google, Force.com, and Azure are all getting better and cheaper by the day. A year from now we won't be having this argument.
    • amen to that ...

      Really! Are a lot of you living under an IT rock? This isn't something in the future; it's happening now. Some of the comments I read are pretty ridiculous. I have to assume two things come into play regarding a lot of the negative comments: (1) Being uninformed. (2) Being afraid that the model your job / business is built on is becoming a dinosaur (and fast). Keep your heads in the sand a little longer.
      • No

        I just don't want to end up bankrupt or sitting in prison for several years...
    • You missed a very serious con.

      Overall user experience. The current Cloud envisions every application running in a web browser. YUCK. Have you used Google Docs? Cheap and simple? Kinda? Quality experience? Only when compared to 1995 native apps.
      • This is very accurate:

        "Only when compared to 1995 native apps."
    • Con - legal

      One of the biggest contra points to the Cloud at the moment is the law. It isn't set up to deal with it, especially if you do business outside of the USA.

      Cloud companies with an office (not even a data center) in the USA are beholden to the Patriot Act and shortly CISPA. That means that upon request, they have to hand over your data - under PA or with an NSA Letter a warrant isn't even required. Additionally, the cloud provider is not allowed to inform you that they have just compromised your entire corporate data.

      Switch that around, outside the USA, there are strong legal protections to personally identifiable data held in electronic form. For instance, in Europe, you cannot hand that data over to a 3rd party (US Government) without a local warrant (issued in the EU) or by obtaining written permission from each affected person.

      As the cloud provider is not allowed to inform you, that they have illegally compromised your data, the first thing you'll know about it is when you have a civil action suit or receive a heavy fine (up to 10% of global revenue) for "allowing" the data to leak.
      • In or Out of USA?

        Actually, there are worse things that having your data in a US database that is subject to the PATRIOT Act: having it in countries that lack BOTH a government unwilling to violate your privacy, AND a government unwilling or unable to stop PRIVATE crooks from violating it, such as Pakistan or Bangladesh or even India.

        The US government will ONLY take control in certain circumstances (and there is currently momentum to oppose the PATRIOT Act from both the left and the right; the right wing supported government spying when THEIR guy was in the White House, but they have mixed emotions about it with the "Kenyan Muslim socialist" having it), but crooks, or even terrorists, could "buy" your data from a corrupt database center in a nation without ANY significant regulation, or means to enforce it. One possibility is to get enough information on about a hundred US citizens who look like the population of the terrorists' country to enable each of a hundred terrorists to claim to have "lost" their passport and get a new one from the US Embassy in the name of a citizen victim. Then once they safely "return home" they would do their dirty work under the false ID of the victim, who unknowingly both supports financially, and is framed to take the blame for, the next terrorist attack.

        So yes, there is a possibility of the US government stealing your data (or more accurately the data about your customers, which belongs to them), but in some cloud providing nations an even bigger risk of someone even MORE unfriendly getting it. And even in US and European Union countries, there is always the risk of hacking by "plain old" criminals.

        In addition, the database could be effectively DESTROYED on the cloud provider's servers, making it unavailable until a backup is restored, and it might as well be gone (for an indefinite period) if either THEIR internet connection or YOURS goes down. Back in the old mainframe days, when dedicated wiring by the phone company connected hosts to the network of dumb, slow terminals in a company (e.g. airline reservation and ticketing systems, truck dispatching, or just interoffice memos and "emails"), critical systems also had a shadow network of dial-up modems that could be used in case of outages of the hard-wired network to create temporary connections to the host. Since the internet is now the only compatible connection, what good is dial-up? In many small and medium companies, and even some large ones, the landline phone has been replaced by VoIP! I remember having callers whose DSL connections on a line for which the landline number was disabled to save money having to use a cell phone to call in for support because their phones were down due to the DSL being down! And some of them had no PC available to debug the DSL because they were using it ONLY for the phone replacement! They don't mention that in the V***** commercials.

        Just as President Reagan said to President Gorbachev "trust but verify", and just as test pilots have parachutes, use the cloud but have a backup CONNECTION and a backup COPY of your data under YOUR control.
    • Add afew

      -Service outage
      Should be added to the list of cons.

      I assure you even BMW wouldn't want its information on Google's servers, even though they don't work in the same industry.
      Ehsan Irani
      • Amazon EC2 has being hacked multiple times

        and was found to be running multiple bootnets in recent days.

        So "security" is also missing from the list.