The Cloudy Future of Corporate IT

by Andrew McAfee on August 21, 2009

As cloud computing gathers momentum in the business world, so too do its critiques and detractors. They tell us that the cloud is insufficiently secure, robust, and stable. That it might suffice for lower-level activities like email and calendaring, but won’t handle chores that require massive computational power, transaction volumes, or bandwidth. That it’s not well suited for a truly mobile worker. That cloud applications are trivialized versions of desktop and client-server ones. That to give up local control over data, software, and infrastructure is to strike a Faustian bargain — one that will someday bring grief. That the full-fledged desktop computer and self-managed data center are here to stay because real industrial strength corporate computing require them. That the cost savings available from the cloud are minimal or nonexistent. And so on…

None of these claims is ridiculous. Each of them, in fact, has a lot of merit. But how many of them have lasting merit? As I listen to most critiques of the cloud I’m not sure if they’re addressing the cloud as it exists now, or the cloud as it will ever exist.

I was reminded again recently of how critically important this distinction is as I was doing background reading for the book I’m digging in on (book #1, on Enterprise 2.0, comes out later this year). This book will be about IT’s impact on business and competition. To better understand this phenomenon I’ve been learning about previous technologies that were a big deal for the economy. Electricity is one of the most obvious of these, and anyone hoping to understand the digitization of business will learn a lot of lessons from the electrification of American manufacturing.

I came across a wonderful 1983 paper by Warren D. Devine, Jr. in the Journal of Economic History called “From Shafts to Wires: Historical Perspective on Electrification.” Devine combed through the contemporaneous business and technology press to learn what ‘experts’ were saying as manufacturing switched over from steam to electrical power, a process that took about 50 years to complete.

Three main points stood out for me as I read this paper with cloud computing in mind:

The real impact of the new technology was not apparent right away. Electrical power didn’t just save costs or make factories a bit more efficient. It allowed radically new designs and approaches. Prior to the advent of electricity, factories were powered by a single big steam engine. This power source drove all the machines in the plant via a complicated system of shafts and belts, an arrangement know as ‘line drive’.

Electric motors were first used as simple replacements for steam power, and factory layouts remained unchanged. But engineers and designers eventually realized that electric motors could be shrunk down and used to power smaller groups of machine in a configuration called ‘group drive.’ This reduced power losses from friction, let parts of of the factory run independently of each other, and increased uptime. Motors were then shrunk even further and applied to each individual machine; this ‘unit drive’ arrangement is the one we still use today.

The shifts from line drive to group drive and group drive to unit drive opened up new possibilities for reconfiguring factories, allowing them to become bigger, more efficient, more flexible, more robust, etc.. US manufacturing became enormously more productive in the first decades of the 20th century. This was not so much because electrification let motors become better, but because it let factories become better.

This happy result was not obvious at the dawn of the electric era, when electricity was seen just as a replacement for previous power sources. Devine includes a couple great quotes to this effect. Prof. F.B. Crocker of Columbia wrote in 1901 that:

There were many factories which introduced electric power because we engaged to save from 20 to 60 percent of their coal bills; but such savings as these are not what has caused the tremendous activity in electric power equipment that is today spreading all over this country . .. those who first introduced electric power on this basis found that they were making other savings than those that had been promised, which might be called indirect savings.

He cited many types of ‘indirect saving.’ For example, electrification eliminated the need for overhead shafts. With them gone, overhead cranes could be installed to carry heavy loads. In 1895, Crocker said:

I do not think any of us rightly conceive of the great convenience and rapidity of work that is coming from the handling of our… loads by [cranes]… this is going to be one of the direct results of the clear headroom brought about by the use of motors.

By 1912, Devine points out, writers about factories considered the value of cranes “so generally recognized as to require no comment.”

The transition to full exploitation of the new technology was long, but inevitable. Electrification of industry in the US began around 1883 and continued for more than half a century. Devine shows that as late as 1930 nearly 20% of factory power was still supplied by steam engines; he also notes while line drive configurations became rare after World War I, they persisted in some old factories into the 1960s.

Full electrification took a long time for a couple reasons. For one thing, it was often cheaper to continue running an old factory the old way than it was to build a new electric one, or even to replace a steam engine with an electric motor. Electrification also required the development of power generating utilities and a transmission grid, both of which took time. It also took time to spread the word and the knowledge about the new technology throughout different regions and industries.

Despite these impediments, though, the march of electrification was inexorable. In most sectors a factory that continued to rely on steam power, or on line drive or group drive configurations, would simply find itself unable to compete with an electrified rival using motors and unit drive. The factory using the legacy technology would be too much less productive, efficient, and agile to stay in business.

At the turn of the 20th century electric motors and unit drive seemed the stuff of science fiction; Crocker wrote in 1895 of “the extreme view… that a motor should be applied to every tool.” But by the 1920s this extreme had become the norm.

There were detractors and skeptics about the new technology throughout the transition. In 1891, Dr. Louis Bell presented the results of his analysis showing that electricity made sense in situations where only small amounts of power were required, but steam was best otherwise. And experts argued about the right answer for powering and configuring factories throughout the following decades. As Devine writes:

… the merits of driving machines in groups or driving them individually were discussed in the technical literature throughout the first quarter of the twentieth century. Between 1895 and 1904, this subject was vigorously debated in meetings of technical societies; neither technique could be said to be best in all cases… And, over 20 years later, group drive was still being strongly recommended for many applications… Two textbooks printed in 1928… make it clear that there were many situations in which group drive was justified.

The people arguing for steam power, line drive, and group drive may not have been wrong about the ‘right’ answer at their particular point in time, but I’m left with the impression that they were concentrating on a moment and ignoring a trend. In other words they were making simple static calculations and failing to take into account that technical progress would, within a few short years, cause the same calculations to come out very differently.

This short-sightedness matters because factory owners made long-term bets at the time they built their plants. If they built for steam power or line drive or group drive they were constrained by these choices, and often unable to take full advantage of electrification even when they woke up to its power. A factory configured for electricity and unit drive from the get-go, in contrast, could take full advantage of all the direct and indirect benefits of electrification as they presented themselves over time.

I hope my purpose here is clear. I’m not bringing up this chapter in business history just because it’s interesting (to me, at least). I’m bringing it up because the analogies between electrification and cloud computing are very tight ones, and are highly relevant for anyone trying to understand how the corporate cloud is going to play out.

So based in part on what I’ve learned about electrification, here’s are my predictions about the move by businesses to cloud computing: It’s going to be a long transition, but an inexorable one. There will be skeptics all along the way, many of whom will have objections and arguments that are valid at particular points in time, but not over longer periods of time. And the longer time periods are the ones that matter, because the move to the cloud is going to be accompanied by benefits that we can’t yet perceive, but that will be substantial.

I believe that what Nick Carr calls The Big Switch is underway. Business computing is in the process of moving to the cloud just as surely as factories moved to electrification about a hundred years ago. We’re at the early stages of this process and it will take a long time to unfold, but it will happen. And business owners who ignore this transition or choose not to participate in it will eventually find themselves constrained and left behind to the same extent as factory owners who sat out electrification.

I know that at present the answer to the question “Should we move this task / process / application to the cloud?” is usually “Well, it depends.” But that’s only because we’re at the relatively early stages of the long transition period. The day will come, probably sooner than many of us think, when the answer to that question becomes simply “Yes.” (or, more likely, something like “Well, duh.”).

There’s clearly a great deal more to say about cloud computing and the transition we’re now going through, and I’ll be writing more about it here and elsewhere. I wanted to kick this topic by looking at a relevant previous period in business history because as Mark Twain said “The past may not repeat itself, but it sure does rhyme.”

What do you think? Do you think the move to the cloud is inevitable? If not, what are the best arguments you’ve heard for why business computing will remain locally owned and operated? And do you think the example of the electrification of industry is a useful one for understanding how cloud computing will play out? Leave a comment, please, and let us know.

  • digiphile

    Great post, professor. In answering your questions, I'll add the caveat that I'm an IT journalist, not a researcher, and that my job is generally to report, not prognosticate.

    Do I think the example of the electrification of industry is a useful one for understanding how cloud computing will play out? Absolutely.The parallel to electrification isn't a novel one, as you acknowledge in your nod to Nick Carr's work, but the paper you dissect and use as a parallel was extremely useful in thinking through comparing the impact of the seminal technology of the industrial to an emerging trend in the information age.

    “Do you think the move to the cloud is inevitable?” It depends whether it's public or private. Not every industry or business can move data to a 3rd party vendor, due to regulatory constraints. Certain business have to go there to compete. Governments, like the commonwealth of Massachussetts or the Feds, may simply build their own data centers to leverage the cloud.

    “If not, what are the best arguments you’ve heard for why business computing will remain locally owned and operated?”
    Privacy. Jonathan Zittrain's scholarship, amongst others, has been useful in thinking through those issues. So too has been talking with state CIOs; they can't risk the information of citizens in third-party server, where they lose control of the data. Again, the distinction between private and public cloud is useful. Any discussion of business computing need to acknowledge virtualization and the rise of the virtual machines, where new workstations and test environments can be spun up on-demand. That may well be powered through cloud computing; the data center itself may simply be in-house. Privacy and yes, compliance concerns will keep coming up in the near future.

    It's worth noting that one of the largest cloud infrastructures in the world is operating by the DoD, as we both learned at the MIT CIO Symposium. If they can find a way to leverage the technologies at play now, it's a good bit that other organizations will follow.

  • cpswan

    I'm CTO of a startup company, and proud of the fact that we don't have a server. We still use a few local productivity applications on our netbooks and laptops (there are no desktops), but all of the heavy lifting is done using SaaS applications. If I ever did need a server for something then I'm pretty sure I'd rent one by the hour from an IaaS provider.

    For me one of the key benefits of cloud computing is that it enables us (as a small company) to punch above our weight. Not too many years ago you’d need a serious outlay on hardware, software and staff to care and feed it in order to have fairly basic services like email and a shared calendar. ERP, CRM and other capabilities were beyond the reach of the little guy because they were too complex and costly. That’s no longer the case – SaaS delivers better capability than many enterprises have deployed, and at a fraction of the cost. What’s missing is the customisation, but I can live without that (and I can also see savvy enterprises running away from customising packages to their baroque processes and instead installing ‘Plain Jane’ software and using the best of breed process embodied in there).

    Over time I hope that the company I work for will grow, but I can't foresee the day when we would say 'we've hit xxx employees, it's time to ditch this SaaS stuff and buy some servers and an ERP package'. That's just not going to happen. What we need from the cloud will grow with us. Big companies fail, small companies grow into big companies, and that's why the move to cloud is inevitable. How many of those companies with their own generators and VPs of Electricity are still around today?

    The more subtle and interesting question is to what extent using specious arguments against adopting cloud computing sooner rather than later will lead to the failure of some of the big companies out there?

  • http://DougCornelius.com Doug Cornelius

    I see three factors that are key to moving towards the cloud.

    The first factor is mobility. As more and more workers access their company systems from outside their offices, cloud computing makes more and more sense. When you are checking you email in your pajamas before breakfast in the morning, you don;t know if your email system is in your office space or in the cloud.

    The second is the anti-cloud company: Microsoft. Its latest operating system and desktop applications are causing CIOs to re-think how they deal with applications. I know of many companies that just moved to Office 2003 and resistant to move to Office 2007 because of the many changes.

    Similarly, many (most) corporate desktops run on Windows XP. I just heard that there is no straight migration path from XP to the new Windows 7.

    All of this is making managing the desktop a big pain on the neck, that adds little strategic value to the firms. The though of having to make radical changes to the desktop computers makes cloud computing very enticing.

    The third factor is the compliance and regulatory concerns. Everyone focuses on the security risks.

    When you look at the data breaches, they mostly fall into two camps: the system break-in and the lost laptop. Cloud computing should remove the lost laptop breaches. The data is not on the laptop its in the cloud.

    What the cloud providers need to show is that their services are just as secure as the company systems. Regulators also need to recognize this.

  • saqibali

    As Jeff Bezos said:

    “Utility computing is Web 2.0’s version of rocket fuel. ‘You don’t generate your own electricity,’ Bezos says. ‘Why generate your own computing?’ The forces driving online apps — internet bandwidth and reliability — also mean that, in terms of data per dollar, servers in your closet or co-location facility can’t compete with industrial-scale bits piped in from hundreds, even thousands, of miles away.”

  • http://sayanghosh.spaces.live.com/ Sayan Ghosh

    Isn’t it going to be a faster transformation than the electricity parallel …. just owing to the age where things shift faster? While the basic premise of the shift is extremely valid, its just the burning pace of change of today which would possibly see cloud computing to pan out and become ubiquitous in very short time. A very intriguing parallel…and impeccable thought process and analysis. Loved this.

  • http://twitter.com/rotkapchen Paula Thornton

    Actually, I find the cloud distinctions/discussions funny. Maybe people don't realize how much Cloud 0.0 already takes place.

    I also find it strange when companies insist on having technology 'in house', when they outsource all their technology. It's all a function of 'control' — which isn't real.

    EDS holds the largest (well did at one time) number of Microsoft licenses (they're regularly contracted by major companies to image, distribute and support all company PC's — not to mention use their collective buying power for same — EDS's biggest department: Supply Chain). In February 2007 when Balmer started talking to the analysts about their strategic direction toward 'the cloud', I saw tremendous opportunity for Microsoft to bypass the questionable value-add EDS provided (as an employee, we got the same 'outsourced' support as clients and it was the worst I've experienced at any major company).

    But then, we could readily talk about the same re: human capital. EDS is also the IT cloud for many company's staff. While EDS shows a large number of 'employees' (137,124 in 2008), a good majority of those are actually former employees of the companies they still work for, sitting in their same desk — only the paycheck has changed.

    It's all just a matter of who's doing the brokering.

  • Stefan63atIBM

    I do completely agree, that Cloud Computing is going to change the IT world in a radical way. We are in the very early stages of getting a clue of the change. Still pur brain sticks with on-premise solutions, not considering Cloud-based solutions. Cloud computing is a challenge, but is an opportunity, too. Companies can gain very fast benefits from Cloud solutions – not only the big ones, but in particular SMB.

    https://www-951.ibm.com/blogs/Stefan_Pfeiffer/e

  • http://twitter.com/pressearbeit J. Wittkewitz

    It is somewhat strange, that we don't read about the grid anymore. If we would do I would agree to the analogy of electrification. Cloud computing is delivering services like web services and SOA we heard of some years ago. But these things have nearly nothing in common with electrification. There is no frying, cooling or heating coming out of the power socket. This is what makes the real difference: the power socket is a universal energy bus whereas the cloud computing is the same old stuff we know since decades: An area where the same old players can do their same old jobs in argueing about standards and interfaces and the like. EC2 and S3 are more likely offers in grid computing and that is the reason why they are as successfull as they are. Salesforce on the other hand is successfll because all Siebels and Navisions failed in the ease of use. salesforce was a smart alternative because controllers could calculate the price mor precisely than ever. No new license modells, no new software generation w/ new personnel capable of delivering services and maintenance. If the big players enter the cloud computing market we will see all this stuff coming into the web because they earn a huge amount of their revenues w/ all this stuff, that drives admins, controllers and users crazy.

  • http://claimid.com/bgrubin Benjamin Grubin

    In response to the previous commenter Mr. Wittkewitz, I think the reason we don't hear about the “grid” anymore is that with the grid metaphor came a strong association with “distributed” computing, which was the ability to break up a single problem and run it across many different individual systems, also known as massively parallel computing. The grid relied on many systems, all alike, capable of running one small part of an overall problem. Unfortunately the applicability of the grid model to the needs of the enterprise are usually quite limited, with instead scientific computing being the largest target for such efforts.

    The “cloud” model is fundamentally different in three ways:

    First, the cloud model doesn't focus on difficult-to-solve problems, it focuses on relatively well-understood problems. Most organizations are quite familiar with how to conceptualize, engineer, build, and deploy a web service, or a financial application, or an HR database. The cloud lets you solve these very well understood, tractable problems faster and cheaper.

    Second, the cloud model leverages economies of scale on a global scale. These same economies on a regional scale were the ones that drove the creation of data centers in the late 90s: cooling and power are much cheaper when you can architect a data center efficiently, and the need for very fast communication between different applications means putting those systems near to each other meant cost savings in high-speed WAN links.

    Third, the cloud model seeks to make IT costs far more predictable and flexible. Instead of having to budget capital expenses to build infrastructure for each individual project, the cloud model allows you to attach a predictable expense stream associated with that service. The expense stream can scale up (or, more importantly, down!) along with business levels. When business levels inevitably fall, you are not left with idle capacity that you are still forced to depreciate, cool, and power.

    It is for these reasons that the electrification model is apt. Infrastructure as a service truly means not having to generate your own electricity, but instead paying for it as you use it, and that's the part of the electricity model that makes sense for most businesses. Unfortunately grid computing was predicated on building new, intelligent services that were able to utilize the new grid computing resource. Cloud instead lets you leverage all of the grid's gains with no changes to your services or applications: literally, plugging in your existing business services to someone elses compute resources.

    Finally, I think it's important to note that the enabler of this change is virtualization technology, which has matured dramatically since the late 90s/early 2000s when the grid computing concept hit the scene. It is highly efficient hypervisor-based virtualization technology that allows you to run your unmodified business service like a web server, database server, or application server, in a faraway cloud instead of your local data center. But virtualization's advantages come at a cost: security. What will hold most organizations back from leveraging this incredibly powerful resource in the next five years is the lack of a fundamental security model that accounts for the mobility of services out of the data center (and therefore behind the corporate firewall) and into a cloud, where an organization is forced to “delegate” the security task to an outside organization. The ability to manage, monitor, and secure services running on remote resources is the giant growth area right now, and when these problems begin to be solved in a standardized way, cloud computing will really take off.

    http://www.linkedin.com/in/bgrubin

  • http://twitter.com/pressearbeit J. Wittkewitz

    OK, just saw, that you work as Marketing Manager for Novell….ur arguments make sense in that perspective…

  • http://twitter.com/pressearbeit J. Wittkewitz

    OK, do smarten up teh discussion. What if Mintzberg is right in his synthesis in terms of new forms of organizations? If so, process modelling and planning would become obsolete. And also providing of high specified business applications that virtualized these tayloristic and taylormade processes. What if every element became decentralized and structures would become adhoc groups? Managers would be obsolete too, because they only do the communication between divisions and departments. Without departments there would be no use of all this old strategic planning stuff. I think, that we face the end of the organizational structures of bureaucracy and with it we will loose the need for its virtualization with web services and an IT infrastructure. Social software is the toolbox that replaces management and processes which will overcome most peoples inhibitions. To make this clear, offers like Google Wave will make deal not SAP or Lotus Notes in the cloud.

  • http://twitter.com/MartijnLinssen Martijn Linssen

    Good story on steam versus electricity!

    In my view, however, I'd rather compare desktops to 'unit drive' and Clouds to 'line drive', so it could arguable be used against the call for Cloud computing

    Nonetheless, I like the idea of Cloud computing, as we all have individual experience with it: we have huge x-core processors in our notebooks with way-too-much RAM and video power, that we only occasionally use for the full 100%, e.g. when we're starting up Windows – it's just too bad that we've already paid for it all

    However, in my opinion there are three major obstacles:

    First, it's nice to be in the cloud, just as nice as having outsourced to a new datacentre, but it's a moneydevouring operation to get in or out of it, which adds little value if any at all. So we need what I call “no man's land” clouds where outsourcers, contractors and basically any operator can walk in and out of – past the security of course. If we have to switch from clouds every 5 to 10 years because a contract ends, it will seriously squeeze the fun out of Cloud computing

    Second, it's fun to do business in the Cloud but while you're at it it would even more fun to do business across that same Cloud. I do think that, at least for the first dozen years, Cloud computing will add a certain “static state” to the apps in there, slicing off a good piece of flexibility. Great organisation and control over your apps can make up for that. Governance and service orientation by superb A2A integration will give you a good buffer for those years to come, and give a head start for the enormous spread of B2B and B2C Cloud computing -in my view- will cause

    Third, public clouds are a bridge too far for many. We'll see private clouds, hybrid clouds, virtual clouds, up to the point that one could endlessly debate whether this particular cloud is “really a cloud” or just an ordinary data centre

    But, I agree, we'll get there in the end, but not at that scale. Our current 'unit drives' have become just too relatively cheap to turn into dumb terminals

  • guodan
  • guodan
  • razh

    The real question is not if your IT will be hosted on a cloud or internaly but how your IT is going to provide you a competitive advantage and support your essential innovation. Given that the cloud infrastructure provides reasonable answers to security, data protection and ownership, bandwith, etc. you still remain with the question how to innovate with your business processes to provide value to your customers.

    Having your IT on the cloud doesn't make it more flexible to dynamic requirements, more easy to integrate with some new innovative software, more adaptable to changing business requirements. This challange is and will not be addressed by the cloud architecture. It is addressed by the way you architecture your IT to support your processes regardless where it is hosted and as long we still strugle with the current monolithic and rigid architectures, we will continue to face the same challanges.

    The cloud could probably will become a good infrastructure that might reduce TCO mainly from an hardware / network / infrastructure and operation point of view. Reducing TCO is a real value and as such should not be under estimated. But still the real competitve value doesn't come from just lowering the IT costs, as important that might be. We still have a long way to make our IT architectures open and adaptable enough to cope with the modern competitive landscape, be it on the cloud or elswhere.

  • http://www.ademero.com/products/document-scanning-software/ Dustin

    Great post professor. Thanks for sharing. But it may look funny to most of the people. They may not realize the how much Cloud 0.0 already takes place.

  • Scott S

    Is the move to the cloud inevitable? In my view, the key will be finding the strategic balance between using the Cloud vs. corporate IT infrastructure. It isn't an all or nothing switch. As previously menioned in another comment, a hybrid approach is likely the most realistic model. A key decision point is integration – many of the cloud based services are “silo services” and they do not offer integration (and if they do, security becomes an issue). For this reason, peripheral services will likely be adopted before core business services. Another aspect to look at – what is the target audience of the public cloud? We have already seen high adoption of cloud based services for individuals. For the small business sector, the cloud represents a huge opportunity to keep IT costs under control and allow them to focus on their core competency (which normally isn't managing IT). In the end, I think the cloud has made corporate IT rethink its approach to building infrastructure. It begs the question, what value is corporate IT adding in building out and managing infrastructure? If there is little value, can it be shifted to the cloud?

  • http://www.longnightgown.net/ long nightgown

    This article is very good. Useful to me. I will return to read this article and other article absolute sure. thank you
    For my site long nightgown

  • 123456789sbb

    wedding dresses,wedding gowns,bride dresses,bridesmaids dresses,evening dresses,bridal gowns,flower girl dresses
    Wedding Gowns
    Formal Gowns
    Cocktail Gowns
    Find the wedding dress designer and wedding dress that's right for you! Browse dresses from
    Bridesmaid Gowns
    Evening Gowns
    View our selection of exquisite, handmade gowns and dresses for your wedding
    Wedding Dresses, Wedding Shoes and Wedding Accessories from wedding shop, the UK's finest collection of designer wedding dresses.
    Use the wedding dress and
    cheap wedding
    wedding dresses
    wedding shop

  • lihaoxj16

    tiffany jewelry
    Choose, buy and shop for on sale tiffany jewelry including Tiffany & Co Silver Necklace, Pendants, Bangles, Bracelets, Earrings, Rings and Accessories.
    tiffany co
    Tiffany Jewellery offering bangle Jewellery, bracelet jewelry, eardrop jewelry, necklace jewelry, ring jewelry, finger ring jewelry and earring jewelry
    tiffany
    tiffany and co
    links of london
    links london
    Tiffany Style Silver Jewelry: Rings, Earrings, Necklaces, Bracelets and more Tiffany Jewellery at low prices.

  • http://cloudapproval.wordpress.com/2010/08/09/does-cloud-computing-make-good-business-sense/ Does Cloud Computing Make Good Business Sense? « Von'Victor Valentino Rosenchild – Blog

    [...] from the smaller players, which often use services such as EC2 to provide their own offerings. As Andrew McAfee recently pointed out, the transition to new economic models is long but usually inevitable and cloud computing is no [...]

  • http://andrewmcafee.org/2011/01/jevons-computation-efficicency-hardware-investmen/ 19th Century Economist Reveals Surefire Investment Strategy!

    [...] got an Internet-connected PC and a smartphone, and after datacenters have centralized and moved into the cloud, total US corporate spending on computing gear will taper off, [...]

  • http://www.hlcgroup.net Katie

    And do you think the example of the electrification of industry is a useful one for understanding how cloud computing will play out? Yes since electrification evolve to new ideas and better built ups that cloud will change into more something that is been expected.

  • Anonymous

    This is precisely what has happened with all utilities.  The piece missing was the propensity for large companies to build their own “in-house” utilities until the public utility was reliable enough to be used for manufacturing.   It was used on the public because outages were less critical at home than in production.  Nowadays no one would build their own power station of phone company.  It is the same with the cloud. Hotmail / gmail was originally conceived as good enough for personal mail but not enterprise mail.  It is all inexorably following the same path towards public utility, on demand, pay for what you use.

Previous post:

Next post: