eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
IBM’s moves into cloud computing have been strategically brilliant because of the tremendous risks the cloud poses to the company’s traditional server, software and services business.
“Cloud” means different things to different people and organizations. Google and Yahoo host all of their software and services on the Internet, making them Web-centric cloud players. Microsoft is pursuing its software-plus-services strategy, tying customers’ desktop software to Web-based services.
Customers don’t need a great deal of data center hardware to leverage Google, Yahoo and Microsoft cloud services, but IBM is a different bird. IBM has dabbled in the cloud in a careful way, offering software and hardware combinations that do not cannibalize its existing server and services businesses.
The company’s iDataPlex server, introduced April 23, is essentially a server unit tailored for Web-based computing. On April 28, IBM unveiled cloud-based storage services.
How did IBM get to this point and where is it going to go in the cloud?
Dennis Quan, chief technology officer of high-performance on-demand solutions at IBM, told eWEEK that IBM entered the cloud internally with its IBM Innovation Portal a few years ago, allowing the company’s 355,000 users to access computer resources more quickly instead of having to buy physical hardware.
This approach allowed IBM to “incubate new innovations and let users give feedback” on them, Quan said. Some 20 percent of the projects created in the portal influenced IBM’s products, allowing IBM to shift its cloud focus into the market.
In December 2006, IBM struck an agreement with Google on cloud computing, and in late 2007 it built out three cloud computing centers, leveraging IBM Tivoli management software and Google’s MapReduce parallel computing model on 1,000 servers.
IBM lets students use this gear and the centers as their playgrounds for parallel computing. This appears altruistic on the surface, but it’s ingenious. It means the future university students who embark on this program will be versed in use of IBM and Google gear instead of Microsoft Windows and Live.
IBM then announced Blue Cloud in November 2007, discussing ways it would commercialize these assets and turn them into hardware and software. This will require a lot of virtualization technology, and some customer rollouts are taking place in Ireland and in China.
To recap: IBM’s cloud moves began internally, spread to universities and are now leaking their way into the mainstream in products such as the company’s new iDataPlex.
IBM Has Reasons for Being Careful with the Cloud
Despite all the theorizing about computing services shifting into the cloud, Quan said, “While some enterprise customers want to access compute resources outside of their firewalls, many insurance companies or financial service companies actually want to keep control over their data centers.”
What these companies are doing is looking at Google’s, Facebook’s and MySpace’s data centers and wondering how they can get that scalability within their own data centers. Blue Cloud and the iDataPlex aim to offer that.
By contrast, products and services from Google, Yahoo and Salesforce.com are natively in the cloud as entirely hosted solutions. They use a mishmash of commodity boxes designed to offer impressive scaling and 24/7 uptime.
Pund-IT analyst Charles King said while IBM already has substantial hosted services offerings, he doesn’t see the company “playing in the cloud-hosting place.”
IBM is picking its spots in the industry, for good reason. Its core business segments are servers, software and global services. If the company took its entire software group and hosted it in the cloud for customers-a la Google or Yahoo-it would begin to eradicate its server and global services businesses.
Who would buy an IBM server if IBM is hosting its own software for customers to subscribe to? Who would pay for the global services unit, famed for offering support staffs to take over corporate data centers?
One could argue that IBM could use its surplus of servers and services personnel to support IBM’s software and internal machines, similar to the way Google puts 100 or so people up in its global data centers to run and maintain them. But the company would still have to lay off a lot of people and jettison some of the weaker products.
It could also be argued that IBM’s server and services business hampers its ability to play significantly in the cloud, but, on the other hand, IBM’s understanding in those data center disciplines positions it to be a leader in the space, stronger than Microsoft or even Google.
“They’re proceeding in the way they traditionally proceed, which is, take a look at the market, see where some discreet commercial opportunities are and then ask if there is anything that they can bring to the party that no one else has done,” King said.
IBM is proceeding into the cloud, as it has every new business opportunity, with caution, and defining its own cloud strategy, not letting the media hype shape it. Ultimately, new opportunities will open up for the company as it goes further down the cloud rabbit hole.
Who can predict what those might be?