Clouds Roll In


Published on

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Clouds Roll In

  1. 1. THE CLOUDS ROLL IN An Afterword toThe Big Switch: Rewiring the World, from Edison to Google By Nicholas Carr In association with April 2011 1
  2. 2. A Meeting in Monarch BeachIn September 2010, I traveled to Monarch Beach, California, to attend theInformationWeek 500 Conference, a large annual gathering of CIOs and otherinformation technology executives. My most recent book, The Shallows: What theInternet Is Doing to Our Brains, had come out a couple of months earlier, and FritzNelson, InformationWeek‘s editorial director, had invited me to the event to give a talk onthe book‘s themes. As I sat in on the other sessions, though, it became clear that theconference‘s major thrust was cloud computing, a topic that had been the focus of myprevious book, The Big Switch. Figuring out ways to capitalize on the promise of ―thecloud,‖ while tempering the risks of being an early adopter of the powerful new set oftechnologies, had risen to the very top of the CIO‘s agenda.During a break in the proceedings, I had the pleasure of sitting down for a coffee withAnubhav Saxena, a top executive with HCL, a $5.7 billion IT services company and oneof the conference‘s lead sponsors. Anubhav had been an enthusiastic reader of The BigSwitch, and he was eager to share with me HCL‘s multifaceted plan to deliver the powerof cloud computing to its global clientele. He walked me through a PowerPointpresentation on his laptop, and he gave me an early look at MyCloud, a service thecompany is developing that will provide companies with a simple dashboard for building,managing, and integrating cloud systems, applications, and services from a variety ofproviders. In The Big Switch, I had argued that the creation of intuitive ―cloud interfaces‖would be essential to the broad adoption of cloud computing, and MyCloud was one ofthe most ambitious cloud interfaces I had yet seen.One thing led to another, and Anubhav was soon asking me whether I had given anythought to writing an afterword to The Big Switch, reviewing the enormous progress incloud computing technology and adoption that had taken place since the book waspublished early in 2008. I told him that, having spent the last two years researching andwriting The Shallows, I was eager to revisit the subject of cloud computing. A series of 2
  3. 3. phone calls and meetings ensued, which culminated in HCL‘s agreement to provide mewith financial, logistical, and intellectual support in writing an update to The Big Switch.The afterword you are now reading is the result of that collaboration, and I am pleased todedicate it to Anubhav and his colleagues.“Cloud First”Two months after the InformationWeek conference, on December 9, 2010, the chiefinformation officer of the United States, Vivek Kundra, released a sweeping plan foroverhauling the way the federal government buys and manages information technology.The centerpiece of the plan was the adoption, effective immediately, of what Kundratermed a ―cloud first‖ policy. Noting that the government had long been plagued byredundant and ineffective IT investments, which often ended up ―wasting taxpayerdollars,‖ he argued that a shift to cloud computing would save a great deal of moneywhile also improving the government‘s ability to roll out new and enhanced systemsquickly.1To speed the adoption of the plan, Kundra ordered the IT departments of everygovernment agency to move three major systems into ―the cloud‖ by the summer of2012. At the same time, he announced that the government would use cloud technologies,such as virtualization, to reduce the number of data centers it runs from 2,100 to 1,300,that it would create a marketplace for sharing excess data-center capacity amongagencies, and that it would establish performance, security, and contracting standards forthe purchase of utility-computing services from outside providers.Once fully in place, the ―cloud first‖ policy, Kundra predicted, would transform thegovernment‘s cumbersome and inefficient IT bureaucracy into a streamlined operationable to deliver valuable new services to the American public. ―The Federal Government,‖he wrote, ―will be able to provision services like nimble start-up companies, harnessavailable cloud solutions instead of building systems from scratch, and leverage smartertechnologies that require lower capital outlays. Citizens will be able to interact with 3
  4. 4. government for services via simpler, more intuitive interfaces. IT will open government,providing deep visibility into all operations.‖Kundra‘s plan was remarkable for its scope and ambition. But even more remarkable wasthe fact that the plan provoked little controversy. Indeed, its release was met with acollective shrug from both the public and the IT community. That reaction, or, moreprecisely, lack of reaction, testifies to the sea change in attitudes about cloud computingthat has occurred over the last few years.When The Big Switch was published in January 2008, awareness of the possibility ofproviding data processing and software applications as utility services over a public gridwas limited to a fairly small set of IT specialists, and the term ―cloud computing‖ waslittle known and rarely used. Many IT managers and suppliers, moreover, dismissed theentire idea of the cloud as a pie-in-the-sky dream. Cloud computing, they argued, wouldnot be fast enough, reliable enough, or secure enough to fulfill the needs of largebusinesses and other organizations. Its adoption would be limited to only the mostunsophisticated and undemanding users of information technology.Today, just three years later, the skepticism has largely evaporated. There is still debateabout how broadly the utility model will ultimately be adopted, but most IT vendors,computer engineers, CIOs, and technology pundits now accept, almost as a matter offaith, that the cloud will be a fundamental component of future IT systems. EvenMicrosoft‘s chief executive, Steve Ballmer, once a vocal critic of utility computing, hasbecome a true believer. He said of the cloud in a 2010 speech, ―It‘s the next step, it‘s thenext phase, it‘s the next transition.‖ At Microsoft, he continued, ―for the cloud, we‘re allin.‖2 A few months later, the software giant put an exclamation point on its CEO‘s wordswhen it announced it would spend hundreds of millions of dollars on a global ―cloudpower‖ advertising program, its largest ad campaign ever.A recent survey of 250 big international companies found that more than half of them arealready using cloud services, while another 30 percent are in the process of testing or 4
  5. 5. introducing such services. Only 1 percent of the companies said that they had rejected theuse of cloud computing outright.3 In addition to Microsoft, most other traditional ITsuppliers, including hardware and software makers as well as outsourcers, systemsintegrators, and consultants, are rushing to roll out and promote cloud services, andleading pure-play cloud providers such as, Amazon Web Services,Google, and Workday are rapidly expanding their offerings and ramping up their salesefforts. Many billions of dollars are being invested every year in the construction of clouddata centers and networks, a construction boom that echoes the one which accompaniedthe rise of electric utilities a hundred years ago.But, truth be told, all this frenetic activity and marketing hype is a bit misleading. Thebusiness world, when it comes to the cloud, is far from ―all in.‖ Outside the federalgovernment, cloud-first policies remain rare. While most companies have embraced thetheory of cloud computing, few have placed the cloud at the center of their IT strategy,and the actual adoption of cloud services remains in its infancy. Companies are, to besure, using some cloud services and experimenting with others, but corporate investmentsin cloud computing, while growing at a healthy clip, still represent a small fraction ofoverall IT spending.4 And the investments that are being made tend to be narrowlyfocused on popular software-as-a-service applications. Only about 3 percent of bigbusinesses are tapping into cloud data centers to run virtual servers, and fully 85 percenthave no current plans to use such utility computing infrastructure, according to a 2009survey.5 An equally high percentage say they have no existing plans to use the cloud fordata storage.6Much of the wariness about moving too quickly into the cloud can be traced to the manyuncertainties that continue to surround cloud computing, including issues related tosecurity and privacy, capacity, reliability, liability, data portability, standards, pricing andmetering, and laws and regulations. Such uncertainties are neither unusual norunexpected; similar ones have accompanied the build-out of earlier utility networks aswell as transport and communications systems. Another force slowing the adoption ofcloud computing is inertia. Many companies have made huge investments in in-house 5
  6. 6. data centers and complex software systems and have spent years fine-tuning them. Theyare not going to tear everything out and start from scratch.For large businesses in particular, we are still at the beginning of what promises to be along period of transition to cloud computing. The cloud is revolutionizing businesscomputing, but this will not be an overnight revolution. It is one that will, as I argued inThe Big Switch, play out over the course of at least a decade—and more likely two. Thatdoes not mean, though, that corporate executives and IT professionals should becomplacent. The current transitional period will be marked by myriad advances andsetbacks as well as many upheavals—not just technological but also commercial andsocial. Making the wrong choices about the cloud today could leave an organization at adisadvantage for years to come.From Dedicated to SharedIt is already quite clear that cloud computing is emerging as the dominant newtechnological model for computer systems. The models it supplants, such as, notably, theclient-server model, entailed tight connections between hardware and software andbetween systems and users. Computers and related gear were dedicated to particularapplications, and those applications were dedicated to particular individuals (in the caseof PCs) or groups (in the case of servers). These models were, as a result, highlyfragmented, characterized by redundant assets and investments and low levels of capacityutilization. The cloud model, in contrast, is built on the assumption of sharing, inparticular the pooling of assets to support a diversity of uses and users. Hardware isshared flexibly by applications, and applications (and other resources) are shared flexiblyamong heterogeneous users. As with any utility system, the shared, or ―multi-tenant,‖infrastructure dramatically reduces redundancy, improves utilization, and in generalcapitalizes on scale economies. 6
  7. 7. As defined by the National Institute of Standards and Technology (NIST), the emergingcloud model has five essential characteristics, all of which relate to the flexible sharing ofassets:  On-demand self-service. A consumer can unilaterally provision computing capabilities, such as server time and network storage, as needed automatically without requiring human interaction with each service‘s provider.  Broad network access. Capabilities are available over the network and accessed through standard mechanisms that promote use by heterogeneous thin or thick client platforms (e.g., mobile phones, laptops, and PDAs).  Resource pooling. The provider‘s computing resources are pooled to serve multiple consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to consumer demand. There is a sense of location independence in that the customer generally has no control or knowledge over the exact location of the provided resources but may be able to specify location at a higher level of abstraction (e.g., country, state, or datacenter). Examples of resources include storage, processing, memory, network bandwidth, and virtual machines.  Rapid elasticity. Capabilities can be rapidly and elastically provisioned, in some cases automatically, to quickly scale out and rapidly released to quickly scale in. To the consumer, the capabilities available for provisioning often appear to be unlimited and can be purchased in any quantity at any time.  Measured Service. Cloud systems automatically control and optimize resource use by leveraging a metering capability at some level of abstraction appropriate to the type of service (e.g., storage, processing, bandwidth, and active user accounts). Resource usage can be monitored, controlled, and reported providing transparency for both the provider and consumer of the utilized service.7 7
  8. 8. Along with the cost efficiencies and flexibility benefits that come with sharedinfrastructure and applications, the cloud model benefits users by distancing them fromthe complexity inherent in building and running modern IT systems. The users purchaseand configure IT services through relatively simple interfaces without regard to either thelocation or the architecture of the systems delivering the services. In other words, cloudsystems allow not only the consolidation of physical assets but also the consolidation ofthe expertise required to build and maintain the systems. The consolidation of assets andexpertise in large-scale, industrialized IT plants also spurs rapid innovation in data-centerand systems design. Already, for example, the growth of cloud computing has led toimportant breakthroughs in energy efficiency and cooling—breakthroughs which in timewill disseminate throughout the industry, benefiting all.Because information technology is highly modular, new architectural models do notsimply replace old ones. Rather they build on top of, or at least alongside of, the oldmodels. Companies continued to use mainframes along with client-server systems, andboth those models will continue to be used along with cloud models. Nevertheless,because of its fundamental advantages in cost, flexibility, energy efficiency, and ease ofuse, the cloud model will increasingly be the model of choice for building new systems.As devices and applications are built to take advantage of cloud systems and centraldatabases, moreover, the cloud model will gain further momentum and becomeincreasingly entrenched at the core of modern information processing.With total worldwide spending on information and communication technology nowexceeding $2 trillion annually, according to the World Bank, the shift to the cloud ishaving far-reaching economic implications. The epicenter of the transformation is the ITindustry itself. The industry is being challenged to move from a business model based onselling or licensing expensive high-tech components (and then assisting buyers inmelding the components into purpose-built systems) to a business model based on sellingservices for fees that vary with consumption. This shift promises to make the industrymore capital-intensive, as large IT providers will need to own and operate data-center 8
  9. 9. networks, often global in reach, in order to provide services, as well as more competitive,as companies that once specialized in particular components (Cisco in networking, EMCin storage, AT&T in telecommunications, Oracle in applications, and so forth) increasingbattle head-to-head.The shift should also, in the long run, lead to the further commoditization of many of theindustry‘s outputs, particularly at the infrastructural level. The cut-rate, low-marginpricing of pure-play IT utilities like Amazon Web Services testifies to the pricingpressures that lay in store for many core IT services. Given these characteristics of cloudcomputing, particularly the pressure to achieve scale economies, we will likely see afurther consolidation of many segments of the industry, with success increasingly goingto those companies that prove themselves most adept at building and running multi-tenantsystems and marketing the services the systems generate.The View from the EnterpriseFor large companies, the cloud is typically used today not as a replacement for in-housesystems but as a complement, supplement, and extension to those systems. Software-as-a-service (SaaS) subscriptions provide an easy, low-risk way to extend or modernizestandard corporate applications, from human-resource administration and accounting tosalesforce automation and project management, while also accommodating the mobiledevices and remote collaboration that are essential to many modern business processes.Cloud-based programming platforms (PaaS) allow for the rapid development, testing, androllout of new software applications and features without requiring investments in in-house servers and development tools. And the ability to rent large amounts of storage andserver capacity and other infrastructural services online (IaaS) enables companies to dealeffectively with the kind of short-term computing requirements often involved inconducting research and analysis projects or meeting seasonal spikes in demand. 9
  10. 10. The pharmaceutical giant Eli Lilly has made the cloud a central part of its vast R&Dprogram. The company encourages its scientists to tap into cloud data centers for early-stage number-crunching in analyzing potential new drug compounds. It has developedmore than a dozen templates for cloud use in different research scenarios, allowingscientists to rent virtual servers for pennies an hour with their corporate credit cards. ―Thecost is trivial in some cases for what theyre doing,‖ Eli Lilly CIO Michael Heim said atan IT industry conference in 2010. ―It‘s hard to overestimate the value of letting scientistswork at their own pace.‖ Using the cloud for research has allowed Lilly to cancel plansfor building a third in-house data center for R&D. The savings have been so great that thecompany is now looking to extend its use of the cloud to ongoing computing needs aswell as short-term projects.8 It is also developing cloud management software that willallow it to easily shift jobs among different cloud vendors, reducing the risk of gettinglocked in to one provider‘s services, and it is looking to set up an industry consortium tofacilitate cloud-based collaboration among drug researchers from different companies.One leading U.S. retail chain is looking to the cloud to reach a new generation ofshoppers who expect to gather information and services through their smartphones andother mobile devices—even when they‘re inside a brick-and-mortar store. The company,according to one of its top IT executives, has spent 20 years building a tightly integratedsuite of in-house systems for managing its far-flung operations, including specializedapplications for purchasing, warehousing, shipping, store management, andmerchandising, and starting from scratch with cloud solutions would be not onlyimpractical but foolish. In the retailing industry, the executive explained to me in aninterview, ―if you have something that works, you dont rip and replace it. You can addnew things, you can enhance things, but the idea of pitching out the baby with the bathwater is never going to be an option in retail.‖At the same time, he emphasized that the way companies think about IT has ―changeddrastically‖ over the past three years, thanks in large measure to the new opportunitiesopened up by cloud computing. Companies no longer look at IT as ―an infrastructure or aplatform or software or an application,‖ he said. ―Everything‘s a service‖—whether it‘s 10
  11. 11. supplied through the cloud or from a private data center. The retailer has alreadyextended its in-house systems with more than 100 applications provided as services byoutside suppliers, and, according to the executive, it will draw on cloud solutions todeliver a range of information services and applications to the mobile devices of the nextgeneration of shoppers. He foresees a variety of web-based apps that will provide extravalue to consumers while also promoting in-store and online purchases.Not all CIOs are taking a ―slow and steady‖ approach to cloud computing. Bob Rudybelieves that the faster his company takes advantage of the benefits of the cloud, the morecompetitive it will become. A former Intel engineer and technology entrepreneur, Rudyhas since 2007 been vice president and chief information officer of Avago Technologies,a $2 billion semiconductor company that was once part of Hewlett-Packard. His view ofIT is hard-nosed—―I want just enough and nothing more,‖ he says—and for the past twoyears he has been aggressively replacing the company‘s in-house systems with cloudalternatives. ―Owning software doesn‘t make sense,‖ he told me during an interview latein 2010, making an exception only for the specialized software that directly supports thecompany‘s core engineering functions. He feels the same way about proprietary datacenters: ―I don‘t want to own a power plant if I can share a grid.‖In 2008, Avago became the first company with over a billion dollars in sales to adopt theGoogle Apps Enterprise suite of productivity tools, including e-mail and calendars. Theshift is saving the company about $1.6 million annually, but the more importantadvantage, according to Rudy, is the gain in organizational flexibility and empowerment.Avago‘s engineers and other employees no longer have to worry about hitting limits intheir in-box capacity, which in these days of rich media can slow their work and impedetheir creativity, and Rudy no longer has to worry about investing precious time andmoney in installing ever greater quantities of new storage capacity to keep up withdemand. In addition to Google Apps, the company has rolled out a series of other majorsoftware-as-a-software applications, including ones from Workday, Taleo, Authoria, andEnlighta. Avago has also moved much of its IT infrastructure into the hands of outsidesuppliers, and is encouraging them to use multitenant systems whenever possible. 11
  12. 12. The bottom-line benefits of Avago‘s cloud-centered strategy have been impressive. Thecompany has already pushed its IT costs as a percentage of revenues down to just 1.6percent, considerably lower than the 4.2 percent average for the high-tech productsindustry, and Rudy says that the figure will soon be less than 1 percent. Rudy‘s onlyfrustration is that he can‘t move even more quickly into the cloud. If he wasn‘t locked into long-term licenses for enterprise-resource-planning applications, he says, he wouldhave been even more aggressive in moving to cloud alternatives. ―We will pursue everycloud service we can,‖ he declares, arguing that the biggest barrier to cloud adoptiontoday is ―fear, uncertainty and doubt‖ spread by IT professionals themselves.What‘s now clear is that there is no one correct recipe for adopting cloud computing, justas there was no one right recipe for adopting mainframe computers, minicomputers, PCs,or client-server systems. Different companies and industries have also employed IT indifferent ways, and those difference will continue to influence the way they adopt and usethe cloud model. But all businesses need to be aware that the risks of moving too slowlymay be as great, or greater, than the risks of moving too quickly. Past practices and sunkcosts are, as always, weak excuses for inaction. As the competitive pressure to gain thecost and performance benefits of cloud computing intensifies, complacency will becomean ever greater liability.A Cloud of One’s OwnWhat may, over the next few years, represent the largest cloud-inspired area ofinvestment for large companies has little to do with the purchase of web-based servicesfrom outside utilities. Rather, it involves the construction of so-called ―private clouds‖—highly virtualized, dedicated data centers that essentially serve as in-house IT utilities. Aprivate cloud can be owned and operated by the company that uses it or, as isincreasingly the case, it can be built and run on the company‘s behalf by an outsourcingor hosting firm or other IT services company. Because the transformation of traditional 12
  13. 13. data centers into private clouds typically requires substantial investments in newhardware and software, to facilitate high degrees of virtualization and automation, theirconstruction has been a boon to many IT vendors. Deutsche Bank predicts thatinvestments in private clouds will reach $20 billion in 2012.The case for a private cloud is often compelling today. Because they typically have muchhigher levels of capacity utilization and scalability than the traditional data centers theyreplace, they can allow a company to gain many of the scale economies and speed andflexibility benefits of the public cloud while avoiding the security concerns, contractingissues, and other uncertainties that currently surround pure utility computing. Just asmany large manufacturers originally constructed their own in-house electricity-generatingstations early in the last century, so many large businesses today are building their ownin-house clouds.FedEx, a long-time IT pacesetter that relies on an array of custom applications tocoordinate time-sensitive shipments around the globe, is in the process of moving to aprivate cloud. It is in the midst of a major effort to retool its core apps to run on astandardized and highly virtualized computing platform, drawing on a common data storeas well as a shared set of data services, such as the provision of a delivery address. Aseach app is updated, it is being moved into a large new cloud data center the company hasconstructed in Colorado Springs. FedEx CIO Rob Carter is convinced that the cloudmodel represents a fundamental breakthrough in corporate IT. ―Whats happening now,‖he recently told InformationWeek, ―is there‘s truly a general-purpose computingenvironment that‘s workload agnostic. You can throw different kinds of workloads on thesame computing server infrastructure.‖ Despite the significant data center and applicationinvestments entailed in building a private cloud, the modernization effort is delivering avery high return on investment, according to Carter. ―For the first time ever,‖ he explains,―you can make investments in a whole new class of technology for about the same priceof just maintaining the base.‖9 13
  14. 14. It‘s important to recognize, however, that dedicated private clouds will in most cases be atransitional technology, a stepping stone on the way to true multi-tenant systems. Atsome point, after all, private clouds will begin to pay diminishing returns; further gainswill require the greater scale that can only come from infrastructure and services that areshared among many companies rather than within just one company. Even today, thedivide between private and public clouds can be blurry. It‘s possible, for instance, tooperate a private cloud as a ―virtual private data center‖ or ―virtual private cloud‖—avirtualized assembly of network, security, storage, and compute resources that isdedicated to a single client but runs on a multi-tenant system. Just as most companiestoday are comfortable using virtual private networks to ensure secure communicationwhile gaining the scale benefits of shared infrastructure, they may well come to embracevirtual private data centers quickly. In the long run, truly private clouds may end up beingoperated only when required to meet legal or regulatory requirements for security orprivacy.It is also likely that we‘ll see the emergence of ―vertical clouds‖—clouds dedicated toparticular industries, or to groups of related government agencies or educationalinstitutions, with resources and performance standards geared to the unique needs ofthose industries. Bob Rudy, the Avago CIO, has begun talks with other semiconductorCIOs in hopes of creating vertical clouds in that industry. Sundeep Reddy, vice presidentand head of IT infrastructure at Toys R Us, believes that vertical clouds may provide away for retail firms to share custom IT infrastructure while also giving them the leewayto maintain competitive differentiation in their applications. It‘s also easy to see howspecialized vertical clouds, providing the benefits of multi-tenancy while ensuring tightercontrols than is possible with purely public clouds, could be attractive to industries withunique data-security requirements such as health care and defense. Vertical clouds couldalso enable the efficient sharing of infrastructure and applications among stategovernments, schools, nonprofits, and other organizations serving similar constituencies.However the tensions between dedicated and multi-tenant resources and services playout, one thing seems very clear: for the foreseeable future, most organizations will 14
  15. 15. operate in a hybrid IT environment, using some combination of traditional in-house ITresources, private or virtual private clouds, vertical clouds, public clouds, and software-as-a-service applications. Developing tools and processes for managing that hybridenvironment and integrating diverse sets of assets, services, and data stores will be a keychallenge for IT departments, and a key competitive battlefield for the IT industry. Onearea particularly ripe for innovation is the creation of user interfaces and dashboards thatallow companies to easily build, integrate, and monitor complex virtualized systems thatdraw on the full set of cloud assets and software-as-a-service offerings. HCL‘s MyCloud,HP‘s CloudSystem, BMC‘s Cloud Lifecycle Management, and CA‘s AppLogic areexamples of the kind of cloud-management platforms that could fundamentally changecorporate IT operations, and further reshape the IT industry, in the near future.In addition to new tools for integrating and managing cloud services, the fulfillment ofthe promise of cloud computing will require cloud providers to join together in adoptingthe kind of clear, coherent, transparent, and measurable performance standards requiredto support contracts, service-level agreements, regulatory compliance, andinteroperability. Much progress remains to be made in this area. A 2010 study of currentcloud-computing contracts, undertaken by the Cloud Legal Project at the Centre forCommercial Law Studies, Queen Mary, University of London, revealed a lack ofconsistency in terms, and also indicated that most cloud providers currently seek to avoid―any warranty of service or acceptance of liability.‖10 The World Economic Forum‘scloud-computing research group has underscored the need for ―more consistent andcomprehensive approaches to accountability for how cloud services are provided‖ as wellas the establishment of clear standards for data portability and interoperability acrosscloud services.Competition among cloud providers may solve these problems, as those vendors thatoffer the clearest terms and standards will likely gain a competitive advantage in themarketplace—forcing other providers to follow suit. But if the industry fails to addressthese critical issues, governments may need to establish licensing or other certificationprograms for cloud providers, guranteeing a basic level of performance, reliability, and 15
  16. 16. security. Establishing and certifying minimum standards in such areas as privacy,intellectual property protection, and data compatibility will be essential to the broad andaccelerated adoption of cloud computing in the future, argues R. Srikrishna, executivevice president and head of infrastructure services at HCL.A Cloud in Every PocketThe cloud turns many old assumptions on their head. Up until recently, if you wanted tosee the cutting edge in information technology, you‘d look at what big corporations weredoing. They were the ones that had the money and the skills needed to build the mostadvanced data centers, procure the latest servers and other gear, and write or buy the mostsophisticated software programs. Today, when it comes to the users of informationtechnology, the cutting edge is found not inside big companies but rather in homes,schools, and startups—those are the places where cloud computing is not only firmly inthe mainstream but has already become, in many instances, the dominant form ofcomputing.Consider your own—or, better yet, your kids‘—experience with personal computing.Five years ago, if you wanted to do something new with your PC, your first instinct wasalmost certainly to go out to a store and a purchase a packaged software application.You‘d bring home the box, slide the CD or DVD into your optical drive, and install theapplication onto your hard drive, making sure it was compatible with your operatingsystem and other applications. Every couple of years you‘d pay for an upgrade and gothrough the same installation process. That‘s no longer the case. Now, when you want todo something new with a computer, you fire up your web browser, hop onto the Net, andfind the data, applications, and services you need. Your first instinct, in other words, is tolook to the cloud—where, more likely than not, you‘ll find what you want, and probablyfor free. 16
  17. 17. The entire Web 2.0 and social networking phenomenon, which has transformed personalcomputing in the last few years, is an outgrowth of cloud computing. A social site likeFacebook, which is now the Web‘s most popular destination as well as an increasinglyattractive platform for online games and other applications, is unthinkable without thecloud. Facebook requires the kind of seamless, large-scale sharing of data andapplications that is only possible with centralized, multi-tenant systems running on theInternet. Most of the popular apps used on iPhones, iPads, Android devices, and othersmartphones and tablets also rely on cloud databases and services for at least part of theirfunctionality, even if they also involve the installation of software code on an internalflash drive. Modern gaming consoles, too, now routinely integrate cloud services servedup from distant data centers. For the first time in the history of personal computing,consumers are today purchasing computers that actually have smaller storage drives thanthe ones they are replacing. Local storage is becoming less important as the cloudbecomes more versatile. It‘s no exaggeration to say that, when it comes to personalcomputing, the ―big switch‖ has already happened.Small companies, particularly entrepreneurial startups, have also often been aggressiveadopters of cloud computing, as have schools, government agencies, and nonprofits. Forthese kinds of organizations, which often have limited capital and tightly constrainedbudgets, the cloud can be a great leveler. By tapping into cloud data centers andsubscribing to Web-based applications, cash-strapped organizations can gain access tomodern, sophisticated IT services that were once available only to big companies withdeep pockets. Indeed, since cloud systems can be continuously updated, they oftenprovide superior capabilities to expensive, installed systems, the upgrading of which isusually costly and time-consuming. As many smaller organizations have found, simplyreplacing an in-house email system with a cheap Web-based alternative can free upconsiderable amounts of money and avoid maintenance headaches and expenses whilealso providing users with far more storage capacity and advanced features than theywould otherwise have had. 17
  18. 18. The common theme here—and it‘s one of the central messages of cloud computing—isthe democratization of data processing. By driving down the cost and increasing theaccessibility of computing power, the cloud continues the long-term trend of making evermore powerful computing resources available to individuals and small groups. If thearrival of the PC meant that everyone had access to a computer, the arrival of cloudcomputing means that everyone has access to an entire data center.This trend also has important implications inside corporations. As the purchase ofapplications and other IT services becomes simpler, faster, and cheaper—as, in otherwords, self-service becomes a reality—end users, such as business units, corporatefunctions, and even individual employees are increasingly purchasing IT servicesdirectly, without routing requests through the IT department. ―Historically,‖ explainsKevin Parikh, CEO of Avasant, an outsourcing advisory firm, ―the buyer of IT services isthe chief information officer of a company. And today the buyer—with the advent ofcloud computing—can oftentimes be . . . someone more directly connected with theservice.‖ As the ―Facebook generation‖ enters the workplace, the expectation of self-service in IT will only grow. A core challenge for IT departments is to facilitate this self-service trend, and the innovation it promises to engender, while also ensuring themaintenance of the controls needed to safeguard corporate data and meet regulatoryrequirements. Eli Lilly‘s use of end-user templates for deploying cloud services willlikely become a common practice in many companies, as will the use of unifiedinterfaces that incorporate management controls.The Path of DisruptionIn the end, the best way to understand the import of cloud computing is through the―disruptive innovation‖ framework laid out by Harvard Business School professorClayton Christensen in his 1997 book The Innovator’s Dilemma. Drawing on a breadth ofresearch into technological advances, Christensen demonstrated that upheavals tend tofollow a pattern. In its early stages, a disruptive technology is characterized by relatively 18
  19. 19. weak performance, which restricts its adoption to companies or individuals with limitedcash, low performance requirements, or both. But the performance of the disruptivetechnology advances quickly, making it attractive to an ever broader set of users.Eventually, the disruptive technology is able to fulfill the needs of even the mostsophisticated and demanding users, at which point it becomes the dominant technology inthe marketplace. Cloud computing is advancing up the performance curve just asChristensen‘s framework predicts, and it seems only a matter of time before it becomesthe IT model of choice not only for individuals and small companies but for the largestcorporations.Beyond the technological changes, the advance of the cloud will mean a realignment ofthe IT work force, with some jobs disappearing, some shifting from users to suppliers,and others becoming more prominent. On the supplier side, we‘ll likely see boomingdemand for the skills required to design and run reliable large-scale, multi-tenantcomputing plants. Expertise in parallel processing, virtualization, energy managementand cooling, security and encryption, high-speed networking and data caching, andrelated fields will be coveted and rewarded. Much software will also need to be written orrewritten to run efficiently on the new infrastructure. In a clear sign of the new laborrequirements, Google and IBM have teamed up to spearhead a major education initiativeaimed at training university students to write programs for cloud systems.On the user side, as the transition to the utility model accelerates in the years to come,we‘ll likely to see a steady decline in jobs related to building and maintaining in-housecomputer systems, while skills in information management and process design andautomation will remain highly valued. We may see as well the rise of a new kind of ITprofessional—a services broker who serves as the interface between cloud services andbusiness units, crafting a flexible portfolio of IT services to meet business needs. Themost aggressive adopters of cloud computing have already experienced a change in themakeup of their IT departments. The size of those departments is shrinking, but thepositions that remain tend to be the more senior, more strategic ones. 19
  20. 20. For years, the knock on corporate IT has been that it is out of sync with business needs—that there is, as the cliché goes, a ―lack of alignment‖ between the IT department and thebusiness units. The weak alignment, it can now be seen, was a symptom of the isolationthat up until recently served as the foundational assumption of corporate IT—isolatedinfrastructure, isolated applications, isolated data, and, all too often, isolated users. Thisassumption of isolation conflicted with the assumption of sharing that underpins businessitself. (The reason business organizations exist is to allow collaboration amongemployees.) By for the first time making sharing the underlying assumption of IT, thecloud promises to finally bring IT and business into alignment—at least for thosecompanies that embrace the cloud‘s promise.CIOs recognize this opportunity, and they‘re embracing it. In fact, the single mostsurprising development in cloud computing over the past three years has been the 180-degree change in the attitudes of forward-looking CIOs. Far from seeing the cloud as athreat, as many originally did, they now view utility-style computing as an opportunity tostrengthen their own roles in their companies by shedding or outsourcing non-strategicactivities and focusing their efforts on core business tasks. Early in 2011, I attended apanel discussion on cloud computing involving a group of prominent CIOs in SiliconValley. They were unanimous in their sense that, as one of them put it, the cloud wasmaking the CIO position ―more relevant‖ than ever. CIOs aren‘t buying all the currenthype about the cloud—they‘re a skeptical bunch—but they now view cloud computing asessential to the future of their companies and their careers.We‘re at the dawn of a new era in business. Just as the last century‘s electric utilitiesspurred the development of thousands of new consumer appliances and services, so thenew computing utilities will shake up many markets and open myriad opportunities forinnovation. We see this transformation playing out not just in IT departments and the ITindustry but across information-intensive industries like media and entertainment.Harnessing the power of the electric grid was the great enterprise of the twentiethcentury. Harnessing the power of the cloud is shaping up to be the great enterprise of thetwenty-first century. 20
  21. 21. About the author: Nicholas Carr is the author of three influential books abouttechnology and its implications: Does IT Matter? Information Technology and theCorrosion of Competitive Advantage (2004), The Big Switch: Rewiring the World, fromEdison to Google (2008), and The Shallows: What the Internet Is Doing to Our Brains(2010). A former executive editor of the Harvard Business Review, he has written formany periodicals, including the Wall Street Journal, New York Times, Financial Times,Atlantic, New Republic, Sloan MIT Management Review, Strategy & Business, andWired. Carr is a member of the Encyclopedia Britannicas editorial board of advisors, ison the steering board of the World Economic Forums cloud computing project, andwrites the popular blog Rough Type.About HCL Technologies: HCL Technologies is a leading global IT services company,working with clients in the areas that impact and redefine the core of their businesses.Since its entry into the global landscape after its IPO in 1999, HCL has focused on―transformational outsourcing,‖ underlined by innovation and value creation, and offersan integrated portfolio of services including software-led IT solutions, ITO andinfrastructure management, engineering services and business process outsourcing. HCLleverages its extensive global delivery centers and network of offices in 31 countries toprovide holistic, multi-service delivery in key industry verticals including FinancialServices, Retail, Media and Publishing, Pharmaceuticals, Manufacturing, ConsumerServices, Public Services and Healthcare. HCL‘s ―Employee First‖ philosophy hashelped HCL recruit and retain the best talent in the industry, growing to over 79,000employees from diverse cultures worldwide. Governed by its core values ‗trust‘,‗transparency‘ and ‗flexibility‘, HCL Technologies is known for delivering consistent ITand business value in mission critical environments.HCL Technologies posted its consolidated revenues of US$3.1 billion (LTM as onDecember 31st, 2010). For more information, log on to 21
  22. 22. References:1 Vivek Kundra, ―25 Point Implementation Plan to Reform Federal InformationTechnology Management,‖ December 9, 2010: Steve Ballmer, ―Cloud Computing,‖ speech at the University of Washington, March 4,2010.3 F5 Networks, ―Cloud Computing Survey: June – July 2009‖: Gartner, ―Gartner Survey Shows Cloud-Computing Services Represents 10 Percent ofSpending on External IT Services in 2010,‖ September 22, 2010: Forrester, ―The State of Emerging Enterprise Hardware: 2009 to 2010,‖ December 1,2009: Forrester, ―Business Users Are Not Ready for Cloud Storage,‖ January 22, 2010: Peter Mell and Tim Grance, ―The NIST Definition of Cloud Computing,‖ October 7,2009: Charles Babcock, ―Eli Lilly Ties Future to Cloud,‖ InformationWeek, September 14,2010: Chris Murphy, ―FedEx CIO Explains the Real Power of the Cloud,‖ InformatioWeek,January 24, 2010: Simon Bradshaw, Christopher Millard, and Ian Walden, ―Contracts for Clouds:Comparison and Analysis of the Terms and Conditions of Cloud Computing Services,‖Queen Mary, University of London, School of Law Legal Studies Research Paper No.63/2010, September 1, 2010. 22