After talking yesterday about grid computing, let's move today to on-demand computing. I didn't comment last week the IBM announcement on its vision of "computing on demand", but Irving Wladawsky-Berger wrote an article which compels me to do it.
In case you don't know, Irving Wladawsky-Berger is VP of technology and strategy at IBM's Server Group. But he's also the guy who will lead the $10 billion on-demand computing initiative unveiled last week. So I guess his remarks are pertinent.
Let's start with a key paragraph about the future of data centers.
Three emerging developments will help address the need for efficiency, flexibility, and simplicity: First, open-grid protocols, which can vastly increase the infrastructure's level of efficiency by sharing resources across highly distributed systems; second, autonomic computing, which is beginning to make the infrastructure more manageable and raise the quality of service; third, on-demand computing, by making applications and information available as an open "utility-like" service, run within the business or outsourced to service providers.
And now, let's focus on the third development: IT on-demand.
As open-grid standards and Web services become prevalent, more business processes, such as CRM or ERP, will become available on demand. The IT staff will simply put these applications out for bid, and choose the best supplier. IT decisions will no longer become architectural issues. They will be made on the basis of business need.
All this has dramatic implications for companies, consumers, and service providers. Increasingly, businesses will regard IT as a utility. And as the focus over time turns more toward the application and the business, rather than the infrastructure, the question will become: "Do I really need all that infrastructure?" Put another way, "Do I want to generate my own electricity, with all the investment in people and equipment that entails? Or am I comfortable leaving the generation of electricity to the experts?" I think the answer to the second question will be a resounding yes.
As IT becomes increasingly standardized, and the open-grid architectural model becomes more widespread, data centers within the enterprise will evolve in one of two directions: They could become smaller if more outsourcing takes place -- in which case the service providers' data centers will grow larger. Or they could grow larger if data centers are consolidated into a few central locations as seems to be happening. Either way, the data center will grow in importance and size. The only question is where it will be located. For the most part, users won't care. They will have what they need, when they need it, largely because CIOs will be shielding them from an incredibly complex infrastructure the way electricity is hidden behind the wall and apparent only in its application.
None of this will happen overnight. But the planets are aligned: Open standards are becoming more prevalent; grids are appearing in the research community and in commercial enterprises; and on-demand services are entering vendor portfolios. IT may indeed be the next utility.
Source: Irving Wladawsky-Berger, for Optimize, November 2002, Issue 13
6:02:01 PM Permalink
|
|
