Is real-time infrastructure too good to be true or a potential new source of reduced cost and increased agility?

IT infrastructure is a critical foundation for business applications, but is expensive, slow to change and hard to manage. Yet it's back on the table as an important source of real business benefits. It seems that real-time infrastructure has brought about this turnaround in the fortunes of "all the boring stuff no one wants to think about".

Real-time infrastructure (RTI) is a very simple idea: Get access to a set of servers and network capacity that grows and shrinks as needs demand - and only pay off the bits that you need at any moment in time. Add complete self-management into the picture and you have the vision for real-time infrastructure. With real-time infrastructure, business demand for IT resources is automatically met, infrastructure is efficient (it is self-managing), fixed costs replace variable pay-per-use costs, and the enterprise can move expensive, rapidly depreciating assets off its balance sheet.

It all sounds just a bit too good to be true doesn't it? There are three strong reasons why real-time infrastructure isn't practical yet: It's almost impossible to make a business case for it; it's not yet technically feasible; and vendor licensing agreements have not kept up with the technology. However, by making carefully placed investments today, it is possible to navigate towards RTI tomorrow.

IT infrastructure is not a commodity, it's a competitive weapon. Enterprises with world-class IT infrastructure are more agile. They have faster time-to-market, greater ability to scale and can integrate mergers and acquisitions faster. Imagine you work for an insurance company that takes 10 to 12 weeks to provision additional capacity on its servers to support some new product. Now imagine you are competing with someone that can implement the same expansion in processing capacity in four to six days. Time, as they say, is money.

RTI is at the peak of the hype cycle. The Gartner hype cycle model exposes the reality associated with real-time infrastructure. Overall, RTI is approaching the peak of inflated expectations - that is, everyone is talking about it, few people understand it, and no one is really doing it. With so much publicity around the concept of real-time infrastructure, and its brother, utility computing, it's no surprise that vendors big and small are talking breathlessly about on-demand and autonomic computing. So be warned. Look beyond the PowerPoint slides because much is missing in these vendors' offerings, and standardization is underdeveloped.

That said, it's not all hype. Although real-time infrastructure technologies are not yet living up to the razzmatazz, some aspects are mature and delivering value. For example, highly available hardware, network load-balancing and storage-area networks are relatively mature technologies, delivering value in many enterprises today. But at the other end of the spectrum, software for dynamically provisioning servers and storage, and automated business policy managers that govern how computing resources are applied to business services, do not currently exist in a production ready form.

Companies with relatively straightforward, homogeneous workloads using internally developed software, like eBay and Google, are in the happy position of being able to virtualize early, without waiting for the general-purpose technology standards. For others, the instability of the technology landscape suggests that CIOs should invest in long-term goals.

Sourcing and licensing are important considerations. Much of the benefit of real-time infrastructure can be captured through internal measures. However, as an organization moves down the path to RTI, considerations for outsourcing change. Internal systems and management become more efficient, but an outsourcing provider's ability to extract economies of scale and efficiently manage their clients' workloads increases.