IBM's enterprise cloud dreams
- TAGS:Blue Cloud, Dennis Quan, IBM
- IT TOPICS:Emerging Technology, Enterprise Software & Services, Internet, SaaS & Cloud Computing, Virtualization
IBM is all over the cloud computing business. Its Blue Cloud program, as described to me this week by Dennis Quan, director of autonomic computing, began less than two years ago and already has established 13 data centers around the globe with more being added every quarter.
Quan points out that when most of us think of cloud computing we imagine a public cloud, such as you get from Amazon.com, Salesforce.com, Google and others. Not public in the sense that your information is out there for all to see, but that the services are out there for all to get.
What's interesting to Quan are private or enterprise clouds, where services are exclusive to a business or government. In fact, as James Hamilton at Microsoft has reported on his blog, the largest users of Amazon's cloud services are not start-up companies and individuals, as you'd expect, but large enterprises. Although these enterprises are using a public service, it does underscore corporate IT interest in exploiting resources in the cloud.
One of the drawbacks to the public cloud, though, is the notion of security. In fact, according to IDC, security tops the list of concerns about cloud computing. That's also one of the top reasons IBM is obliging its customers by investing in private clouds.
So, just what distinguishes an enterprise or private cloud from a public one?
According to Quan, enterprise cloud services are run in data centers in, well, the cloud that are managed by third parties, but also in data centers on company property supervised by corporate IT staff. He says, private cloud services are scalable, growing and shrinking with demand. They need to be accessed via an IP network. And they need to be centrally managed, preferably with Tivoli, of course. Finally, the services need to run in a virtualized environment.
Such an architecture, Quan argues, lends itself to becoming massively scalable beyond the initial private virtualized data center to more external private or even public cloud services based on demand or, I suppose, application choice. Quan believes that dynamic workload management eventually will get so sophisticated that allocating fully secure resources around the globe on the fly to applications on demand will be a breeze.
It's too early to tell if, indeed, that enterprise cloud heaven will come to pass. But it's pleasant to dream that it will.



