An in-depth look at the tech companies jockeying for position as the software industrys leading megatrend takes shape.
Look at the figures for IT spending on cloud computing and youll have to squint to see the skinny percentage. Worldwide dollars directed toward the cloud accounted for a scant four percent of global IT spending in 2008, according to research firm IDC.
But if you gauge levels of vendor excitement about this new model in which software is rented over the Web instead of installed in-house youll realize they sense a once-in-a-lifetime gold rush.
Excitement runs hot. Seemingly every vendor now uses the magic term cloud in their software/hardware product. One of the new netbooks is called a Cloudbook, IBM touts its Blue Cloud, and Dell even attempted to copyright cloud computing (so far unsuccessfully).
Given the tiny four percent spend, are these hyperventilating vendors chasing after pennies? Whats all the deep breathing about?
Top tech giants are looking to the future when IT forecasters do indeed see gold in them hills for cloud vendors. By the year 2012, if you accept IDCs prediction, the cloud will account for a whopping 25 percent of new tech spending. Thats serious money by anyones measure.
The winner in this emerging world will likely claim dominance in the overall tech landscape. That's because cloud computing encompasses most of whats coming: the hyper-growth in connected devices, the surge in real-time data streams, online and mobile commerce, and business use of service-oriented architectures, virtualization and Web 2.0 applications.
The cloud wont just be the dominant software delivery model, it will be an essential part of all consumer and business computing. (Unless, of course, it turns out to be the fad of the early 21st century we never did got those jetpacks they promised us in 1967.)
The $30 billion dollar question is: whos leading the pack into this new era? Whos the big dog that everyone has to keep up with if they expect to survive into the profitable era?
There isnt one, says David Smith, a Gartner analyst. Despite the beckoning treasure the cloud market is still too inchoate to call a leader.
Yet there are top contenders. If you look at the various sub segments, Amazon is clearly an early leader. Salesforce and Google are early leaders, he says.
I do consider Microsoft as, I wouldnt say early, but a current visionary in terms of what theyre looking to offer, he adds. The breadth of Redmonds Azure initiative looks promising, even if parts are still vaporware.
And IBMs Blue Cloud benefits from the companys large enterprise muscle.
The shift wont happen overnight. Most of the top players have unveiled their plans in just the last year or so. And businesses customers, dispirited by the recession, dont want to hear about anything newfangled. Worse, the term cloud has a confusing jumble of meanings and is still seen as marketing voodoo by harried middle managers.
Theres an incredible amount of inertia, Smith notes. Changing something as complex as IT operations is necessarily cumbersome and slow.
Among the potentials of the cloud is that companies will purchase their raw computing power, along with software, from remote datacenters. (This is Nick Carrs Big Switch theory.) The in-house IT department will fade.
Yet companies have invested massively in their IT infrastructures. In-house IT is inextricably linked with business process even as cash-strapped companies look at IT departments as financial albatrosses. In short, classic on-premise IT isnt about to die just yet.
Now, ITs out of the bottle, Smith says. Everybody has it. For better of worse, they know what it is. And youre not going to get it away from them unless they have a really good reason for giving it up.
Two key factors might prompt businesses to trend away from on-premise IT, he says. One is if they can save an enormous amount of money not ten percent, but an enormous amount. Or, if they can do things they couldnt do otherwise. And to some extent thats not really giving it up, but using it for new things.
The five leading cloud contenders Microsoft, Salesforce, Amazon, Google and IBM hope to offer customers those two reasons. As the competition gets serious, heres a look at their strategies:
Next Page: Microsoft: Transformational Changes to Play Catch-Up
Or, go the page that describes a cloud vendor's strategy:
Microsoft: Transformational Changes to Play Catch-Up
Conventional wisdom says that Microsoft needs to worry as customers migrate to the cloud. The software giant makes the bulk of its cash the good old-fashioned way, from pre-installed PCs and shrink-wrapped applications.
Can the 500-pound gorilla maintain its status as the world shifts underneath its feet?
But the company is ready, or at least is in the process of getting ready. In the last few years Redmond has touted its software plus services mantra, which was kind of an either/or strategy. (I interviewed a Microsoft exec about it.) You could certainly buy software in the classic on-premise model, and you could also add in some of their online products. Some of the pieces are substantial like Office Live Small Business and SharePoint Online yet ultimately software plus services looked like a holding pattern until Redmond assembled a full-fledged cloud approach.
In late 2008 Microsoft formally unveiled Azure. In summary, Azure lays the foundation for an integrated online-offline heterogeneous environment, integrated with the Windows OS and the Microsoft development platform.
To say that Azure is ambitious is like saying Los Angeles is little crowded in rush hour. Azure is a sprawling, complex effort that has a cloud-based answer for every question and takes a jab at every competitor. If Azure is successful it will make Microsoft more dominant than ever before, in both the business and consumer markets.
The Azure vision, which ideally will become reality over the next two years, includes:
System infrastructure: Dubbed Windows Azure, this is a cloud-based operating system that offers remote computing power, storage and management services. Its based on a version of Windows Server 2008 and includes the companys free virtualization product Hyper-V (making it a blow against pricey virtualization competitor VMware). Because Windows Azure offers actual computer-processing power from Microsoft datacenters (not just software) it competes with Amazon Web Services and other vendors with physical datacenters.
Application infrastructure services: Known as The Azure Services Platform this is Microsofts software-as-a-service offering for enterprise customers. It includes Microsoft development platform .NET along with business-oriented software tools like SharePoint and Dynamics CRM. This arm of Azure competes with cloud-based application development environments like Salesforces Force.com and Google App Engine.
In a measure of how seriously Microsoft is taking the cloud, Azure even does what is unthinkable for Microsoft: it will be (essentially) platform and OS interoperable, so a non-Windows machine will, in theory, interface seamlessly. Its conceivable that some day an open source business will tap into Azure. Translated: Redmond understands theres dollars in open source and doesnt want to lose a single one, even if it means admitting they wont win every battle.
Perhaps the most eyebrow raising aspect of Microsofts cloud strategy is the massive datacenters the company is building. This is a huge investment. Its facility in San Antonio, for instance, covers an impressive 11 acres. (The one in Chicago is reportedly twice the size.) Preliminary reports indicate the company will build two dozen of these $500 million, 500,000-square-feet facilities.
These mega-centers will ostensibly enable Microsoft to dominate in the brave new world of the cloud, when businesses will buy computing power remotely instead of maintaining in-house data centers. In other words, Microsoft is so committed to the cloud that its spending a royal sum on hardware an unusual course for a software company that has always left hardware to someone else. This shift from selling software alone to selling software and utility computing power is a transformational change in Redmonds approach.
I think theyre taking a big bet on the fact that a lot of enterprise apps are going to move into the cloud they just want to make sure its their cloud, says Jeffery Lindsay, an analyst with Bernstein Research. I think [Microsoft Chief Software Architect] Ray Ozzies got them betting big on it. If hes right theyll do great, and if hes not right then you know
I think Ray Ozzies vision is that they can make a developer environment and then tie that right into the operating system on the computer, he says. This online-offline hybrid will link into Microsoft enterprise libraries through .NET and SQL, with the whole environment powered remotely by Microsofts mega-datacenters. This interlinking of development platform and utility computing will maintain and even strengthen Windows as a top brand, the company hopes.
One skill set of Microsofts that will likely help them lead the cloud and this is critical: Microsoft understands management, says Andi Mann, an analyst with Enterprise Research Associates.
Google doesnt, Amazon doesnt, Salesforce sort of does. In contrast, Microsoft has a full management platform. Theyve got systems set-up, theyve got security management, application management, uptime, availability, recoverability they have management capabilities.
Next Page: Salesforce: An Early Leader Holds its Position
Or, go the page that describes a cloud vendor's strategy:
Salesforce: An Early Leader Holds its Position
The plucky and unabashedly self-promoting Salesforce can claim first-mover status in the software-as-a-service world. Founded in 1999, it gave legitimacy to the once radical idea that enterprise applications would be delivered from a Web site.
One of its rallying cries is The End of Software, that is, applications are now valued only for the services they provide all of which are available online and we should dispense with our troublesome installed apps.
Salesforces sales tracking and account management software (CRM software) has been a clear success. As of 2008 the company claimed north of 50,000 customers. Extending on this, the companys AppExchange, launched in 2005, offers a library of third party software (office productivity, finance, marketing, etc.) that customers can incorporate into their Salesforce mix.
Now its aiming for a much larger goal: Salesforce wants its programming platform to become the default choice for businesses and developers that create applications.
In the same way that programmers who code apps for a mass audience often choose Microsofts .NET, programmers who create cloud-based apps will choose the Salesforce development platform or so Salesforce hopes. The companys application platform, Force.com, was unveiled in September 2007. (The cloud keeps generating new acronyms; Force is referred to as Platform as a Service, or PaaS.)
If the companys development platform were to become the leading choice it would give Salesforce a huge long-term advantage, but actually achieving this will be an astounding feat. As cloud computing moves from emerging idea to a mainstream choice, history suggests products of smaller players dont become the default choice. (Salesforces market cap is $3.3 billion, as opposed to Amazons $22 billion or Googles $102 billion.)
Then again, Salesforces development environment is ideally suited for the clouds future, says Rebecca Wettemann, an analyst with Nucleus Research. Indeed, she points to the company as the ultimate winner in the cloud battle.
Rather than just thinking about the technology, theyve really put in place the whole ecosystem to deliver: developer kits, testing, sand box, developer ecosystem, a way to monetize applications, she says. There are a bunch of things that make a new software developer or even an existing developer very attracted to the Salesforce platform.
And though its a smaller company, Salesforce has what any serious cloud player needs: a network of datacenters. It has datacenters in the U.S. and in May 2008 announced that it would build its first international center, in Singapore.
The companys strategy makes great use of partnerships. Its newest initiative, Service Cloud, leverages its partnership with Google, Facebook and Amazon to enable businesses to create heres the magic phrase, circa 2009 a social networking community to aggregate/explore/influence customer response to their products.
Despite all its success (or maybe because of it) the continual buzz about Salesforce is about which large company will acquire it. The chatter is all speculation, as CEO Marc Benioff maintains the company is not for sale. But even as you read this there are stock traders eyeing Salesforces share price and guessing when itll get snapped up.
Below youll see the traffic for the Salesforce.com Web site, which has remained consistent over the last year (but realize that traffic to the site isnt exactly reflective of app usage):
Next Page: Amazon: B2C Player with Hopes for B2B
Or, go the page that describes a cloud vendor's strategy:
Amazon: B2C Player with Hopes for B2B
Most Web users think of Amazon as the mothership of online retailing, hawking everything from books to perfume to organic peanut butter. (You can even buy a robot duck kit for $19.99.) Yet the e-commerce king is also a cloud pioneer that offered a robust service menu well before most competitors.
The companys cloud strategy grew from its e-commerce roots. As Amazon grew, the number of online retailers connecting to it to sell goods grew as well. Its army of third party retailers is now vast. As it gave small vendors ever more access to its APIs (application program interfaces), it realized that providing Web-based services could be an extra revenue source.
In 2003 CEO Jeff Bezos began developing a business plan around this new idea. Helping him was a big thinker in distributed computing, Werner Vogels (he pens the well-read blog All Things Distributed), who became Amazons CTO in 2004.
Fast forward to 2009 and Amazon Web Services (AWS) boasts a plethora of cloud offerings. The leading components are:
Amazon Elastic Compute Cloud, EC2: virtual servers that customers access over the Web, scaleable as needed.
Amazon Simple Storage Service, S3: a data vault that integrates closely with Amazons SimpleDB, an online database tool.
AWS enables businesses to do upper level tasks such as accessing major applications from Microsoft, Red Hat, Oracle and Sun all over the Internet, on demand. Beyond its core products are myriad peripheral services like Mechanical Turk, which helps manage work among distributed staffers, and Alexa, which tracks Web traffic.
AWS claims 440,000 registered customers, many of which are small businesses and developers. It also touts some big fish clients like Phillips, Nasdaq and the New York Times.
Amazons low-priced cloud tools have earned it a reputation as a kind of entrepreneurs helper. It allows start-ups with only few bucks to leverage a robust technical infrastructure without needing to invest in an IT department. Plenty of SMBs appear far larger than they really are with the help of an AWS back-end.
Clearly, Amazon has dreams of taking over ever more of the IT workload from businesses. It positions itself as a seamless extension of the corporate data center. It can handle an IT departments need for overflow capacity, so-called cloud bursting, or in theory allow an established firm to cut its in-house IT department altogether.
Despite having gained an early lead in the cloud, however, a central question hangs over AWS.
A potential concern is that theyre not really differentiating, Mann says. Microsoft is obviously a major management play, Salesforce is dedicated to a specific application, Google is a very strong infrastructure play with great understanding of a very broad application. Amazon? I dont know. How do they differentiate?
Theyre fundamentally just an infrastructure player or an MSP [managed service provider]. Base infrastructure providers are a dime a dozen and its a very low margin business.
Bernsteins Lindsay echoes Manns sentiment. Amazon probably isnt going all that far in it [cloud]. Its got great utilities; you can get very low-cost computing from its virtual machines. And its got some other services that are kind of cute but good, like Mechanical Turk. But its partly a gee-whiz factor, and partly a way for them to service an online developer bank for people who need to build online stores.
In sum, Amazon is largely offering utilities, Lindsay says. In contrast, the players that will eventually dominate the cloud will offer their own business services and online business processes. Lindsay leans toward Salesforce in that regard.
Indeed, Lindsday doesnt see the cloud as a big revenue source for Amazon. Amazon is a retailer with a $19 billion revenue stream, but if you just look at the very best players [in the cloud space] youll find that their revenues are numbered in the hundreds of millions, he says. The company doesnt break out its cloud revenue separately, but If you look at their revenues from the [cloud] business its probably under $10 million dollars at the moment.
Yet these analysts may be underestimating the clouds division into both a consumer and an enterprise market and Amazon has a profound skill set in the online consumer market.
Theyve taken the worlds second oldest profession, commerce, and put it into the cloud and made it pretty much the way weve always liked to do commerce, says Joshua Greenbaum, analyst with Enterprise Applications Consulting.
The way that Amazon manages the consumer relationship is absolutely unparalleled in its excellence, he says. I expect them to do a really fabulous job in B2C cloud computing.
And for its enterprise play? Amazon wants to be more B2B but I think fundamentally they will devolve toward more of a B2C play.
Next Page: Google: A Giant Work in Progress
Or, go the page that describes a cloud vendor's strategy:
Google: A Giant Work in Progress
Since cloud computing takes place over the Web, a vendor with its own browser holds a natural advantage. Microsoft gains from the ubiquitous Internet Explorer, and the only other competitor so equipped is Google, which launched Chrome in the summer of 2008.
Conversely, one cloud tool that Google has that Microsoft is now just building is sprawling mega-datacenters. Google has many of these facilities, placed strategically around the globe and harnessing the power of countless servers. The company is the king of Internet infrastructure. Its also, of course, the king of Internet service as its search portal aids zillions of users each day. Its technical strength combined with its Web brand name gives it awesome potential in cloud computing.
One thing that Google has going for it is economy of scale, Mann says. I think they have the ability and the understanding of exceedingly large scale environments that theyre going to able to leverage.
They have a very good understanding of how to do load balancing and scalability, geographical balancing, for widely distributed applications. I think that lends strong advantages to a large international Web application presence.
The two-pronged core of its cloud effort is Google Apps, a suite of office tools (calendar, word processing, spreadsheet, etc.), and App Engine, a scaleable development and hosting platform. App Engine, launched in the spring of 2008, is similar to Amazons AWS in that it offers storage, virtual servers and a database. For developers its a pre-built support structure from which to sell their software.
Users can tap into software hosted on App Engine, perhaps accessing software thats custom designed for them by independent developers. And since its hosted by the giant Google the network infrastructure is muscular enough to withstand the heaviest tsunami of user traffic.
That all sounds promising but in reality Googles cloud success has been modest, given the companys vast resources. Although App Engine was launched to considerable fanfare, it has so far largely been underwhelming. Developers grumble that its limited to the Python programming language an odd limitation given Googles fondness for open platforms. (The company says an additional language will be added soon.) Its unclear how many truly high profile applications are being coded on App Engine. Look at Googles own gallery of App Engine success stories and youll see they tout BuddyPoke, Pixverse and CloudStatus while they may be promising applications, theyre not yet Fortune 500 favorites.
Google Apps, too, has a lackluster record. On the plus side, its popular in schools, and in October the District of Columbia became one of its biggest accounts with 38,000 seats. On the other hand, Zoho, a small player whose office apps resemble Googles, recently won the GE account, with its 40,000 seats. The Zoho win was a real poke in the eye for Google, says Lindsay.
Google Apps greatest weakness is its difficulty in attracting lucrative enterprise customers. Google will succeed in the cloud in so far as theres consumer-driven demand for cloud services, Greenbaum says. I dont believe they have an enterprise bone in their body.
Adds Lindsay: Google Apps is losing accounts because what theyre not understanding is that when you have enterprise clients they want very, very specific functionality. And you have to be very responsive otherwise theyll take their business somewhere else.
I think the problem for the Internet-based players, particularly Google, is that they dont have the full enterprise responsiveness and the sales force to service these customers. Microsoft does, IBM does, and new players like Zoho do, and so does Salesforce.com.
I think the way Googles going to go is that theyre going to try and glom into Salesforce and try and learn from them, Lindsay says. (Again, the desirability of Salesforce helps fuel talk of its acquisition.)
Despite the naysaying about Googles cloud prospects, its probable the companys position will grow stronger than it now appears. Cloud computing is a Web-based activity and Googles culture lives, breathes, eats and sleeps the Internet. While the search giant is clearly experiencing growing pains in the cloud, this young and very wealthy company cant be counted out.
Next Page: IBM: Returning to Historic Roots
Or, go the page that describes a cloud vendor's strategy:
IBM: Returning to Historic Roots
The irony about IBM and cloud computing is that, as forward looking as the cloud trend is, it actually plays to IBMs historic strength.
In the 1960s, when IBM was the stand-alone monolith of technology, legions of users tapped into its hulking mainframes. This was the glory of centralized computing much like todays cloud vendors hope users will tap into central datacenters.
The moral of the story seems to be: stay in business long enough and your earliest trends eventually come back in style.
IBMs current entry in the cloud is Blue Cloud, unveiled in late 2007. Blue Cloud enables enterprises to run heavy-duty applications with deep databases over the Internet. The components include grid computing applications, virtualization tools Xen and IBMs own virtualization software PowerVM, open source distributed computing app Hadoop, and Tivoli, IBMs server management software. Its bulked up with plenty of IBM blade servers running Linux and/or IBMs system z mainframes.
Perhaps most potent, given the confusion around this emerging technology, in late 2008 IBM launched a consulting and implementation service for cloud computing, leveraging its longtime strength in IT consulting. The company is also expanding its cloud hardware backbone. In June 2008 the company announced cloud centers in South Africa and China, complementing its existing centers in Ireland and the Netherlands (an early client is the Vietnamese government). In other words, its ambition appears to be to employ a global network of cloud centers.
IBMs Blue Cloud generates less buzz than other vendors initiatives. That may be because, unlike the other competitors, it has no consumer-oriented offering. Or, more likely, IBMs big bucks enterprise customers are still getting used to the idea of the cloud. So far, many of the first brave cloud users have been small businesses and start-ups. For large enterprises, the cloud presents a trove of headaches. The potential pitfalls start with Sarbanes-Oxley compliance, continue on through security for sensitive data, and end up with the dark terror of system downtime a terrible thought for anyone managing a national supply chain. (And cloud vendors were plagued with significant service outages in 2008.)
I think IBM will have a potential play in the very large corporate world because I think thats where theyre going to be strongest, Greenbaum says. There are going to be companies like SAP and Oracle who are going to pick a cloud provider or two on which to host their own cloud services I dont believe that either of them is going to build the infrastructure to provide it.
Adds Bernstein: IBM is the natural for this kind of thing, thats their legacy, their heritage is time-shared computing. If anyone has the pedigree for this it would be them. Now whether they will is another question.
One strategic choice that may help: IBMs cloud plans are built on open source and open standards. Given that the cloud ethos is about interoperability between a vast labyrinth of networks and applications, its likely that open source will age better than a proprietary solution, particularly for a vendor with international partners.
With these open standards in mind, IBM has partnered with Google to provide university computer science students the open source development tools, hardware and services to turn out tomorrows cloud programmers. Reports indicate the two companies have pledged $30 million over two years to the program.
Indeed, IBM and Google appear to be having a bit of a love-fest when it comes to the cloud. (And perhaps in other areas; IBM processes accounts payable for Google.) IBM CEO Sam Palmisano told the New York Times the partnership works well because Google focuses on the consumer market while IBM is so corporate focused. Were more complementary than anything else, Palmisano told the newspaper. We dont really collide in the marketplace.
In May 2008 Palmisano and Google CEO Eric Schmidt appeared onstage at a conference in Los Angeles and announced that they would jointly develop a worldwide cloud network. Presumably the pieces each company has built singly will be included in this effort, but the exact nature of their collaboration hasnt been revealed. One possibility: Googles global network paired with IBMs traditional IT management skills.
Whatever the details, with two giants like this combining forces its probable theyll create an alliance that will profoundly shape the cloud, and theyll have as good a chance as any player to actually lead in the long term.
Heres a chart that provides an overall comparative look at the current vendors in cloud computing:
Go the page that describes a cloud vendor's strategy: