For IT directors, the new data centre road map is clear: Head toward an all-IP network, complete with advanced IP applications such as real-time collaboration.
Already, companies are moving along the stepping-stones to this all-IP place. Many have implemented voice and video over IP, and are turning out frame relay networks and other private links in favour of pure IP connections, such as provided via MPLS-based networks.
For network leaders plotting their courses to the advanced IP paradise, the goal is to achieve real-time collaboration with advanced IP applications and an all-IP network.
In the third of our case studies, we hear from Vijay Sankaran, manager of enterprise technology for Ford's IT group at the car giant's headquarters in Michigan, and Chuck Yoke, director of business solutions engineering for a corporate network in Denver.
A bumpy ride
All-IP is a foregone conclusion at Ford: "Really it's absolutely inevitable," says Sankaran. But he adds that all the pieces to the puzzle must be in place first.
Car-makers, for example, have to "apply discipline of code development to IP-based devices" being embedded in vehicles, Sankaran says. Such devices send information on a vehicle's "health" to manufacturers and service providers over the IP infrastructure. The explosion of these IP-based embedded devices is putting tremendous pressure on IT groups. Software upgrades are difficult once a car has been sold, and defects due to software problems are on the rise across the auto industry, Sankaran notes.
"We grew up in a component-based world. Now we have to think about integration of these IP-based devices within products," he says.
Data management is another critical issue, Sankaran says. "With all these devices communicating, how are you going to manage that information? How will you capture that data and store it?"
While Sankaran worries about the management of software and data across the IP network, Chuck Yoke says the connection layer itself is what will stand in the way of a network utopia.
"In an advanced IP paradise, I'd have true real-time collaboration from anywhere. For instance, I'd be able to use my BlackBerry to pull up an MS Word document and edit it or participate in a whiteboard session remotely," he says. "I'd never do that today because of the speed of my wireless connection. That's one of the biggest technology problems. You can put IP everywhere, but getting it into a usable infrastructure - that's going to be a challenge."
The problem lies with the carriers and their inability to guarantee a high-quality, end-to-end connection, Yoke says. "Carriers are going to have to set a standard for how quality of service will be delivered end to end in a ubiquitous way - that's not a technology issue, that's a political issue," he says.
As for the advanced IP applications, Yoke doesn't foresee problems. "The easiest part of this is the application layer. Web-based applications and Web services will enable us to put applications in an IP-anywhere environment," he says.
But enterprises should not expect to see the returns on IP everywhere overnight, Yoke cautions. "The value will come when you can eventually implement centralised and standardised infrastructure that will not only entail IP-related traffic like voice and data, but much more," he says. "By collapsing these disparate infrastructures and reducing headcount, you will drive the technology value."