March 7, 2011 - 5:34 P.M.
Robert L. Mitchell

By 2020, businesses that fail to pare their legacy architecture may find their core businesses disrupted by smaller, nimbler companies who have built on SaaS and cloud computing from the ground up. "They will be unencumbered by a legacy of complexity and costly solutions and will be narrowly focused on their core competencies," said Frank Wander, senior vice president and CIO at Guardian Life Insurance Co.

And there you'll be with your client/server architectures and your mainframe. Good luck.

That's the picture Wander painted during his state of IT in 2020 presentation at the Computerworld Premier 100 conference today.

Wander sees a gradual migration from applications and platforms in the cloud to data centers in the cloud. "You won't even need to buy a data center -- you will just log in and buy a seat. Think about how much more cost effective that is," he said.

Because of that, competitors may be able to build a corporation much more quickly. "Just imagine how quick and powerful and nimble they will be, and if you haven't uncluttered your environment -- wow," he said. "That's the challenge for legacy companies that haven't invested in eliminating the baggage they're carrying on the journey."

Guardian Life is moving fast. "We are whittling that down as quickly as we can to invest in our strategies," he said. To that end, the insurer is in the process of consolidating six data centers to two -- one in house and one in a hosted facility. "We will have a pod and put the data center in there on a short-term lease," he said. He's also moving the company from Unix to Linux "in a very large way," has already moved 15 back-end applications to SaaS and is focused now on transitioning e-mail, IT services and HR into the cloud as well

Here's the rub: Technology has a long tail going all the way back to the mainframe, and it's unlikely that even an innovative company like Guardian Life can make a complete transition.

Even Wander doesn't expect to go that far. But should he?

Can companies survive in the new world if they drag that mainframe tail with them into the future? Wander thinks so, but only if everything that can possibly be mined from legacy infrastructure -- mainframe, client server -- even personal computing -- is migrated. The fact is that legacy mainframe code is so large, and the cost of reengineering so enormous -- transitioning just one core application can range from the hundreds of millions to more than $1 billion -- that companies simply can't make a business case to migrate.

Gregory Schwartz, CIO for the financial services company USAA and a Premier 100 honoree, said he doubts that upstart competitors can gain an advantage just because their architecture runs on a cloud-based mid-tier or massively parallel processing virtual data center instead of on his 12 mainframes. For customers, the front end is already modernized, while on the back-end, the cost of the mainframe has continuously dropped. Even business process rules have been externalized to the mainframe. If the application performs and it's cost effective, he said, it doesn't matter if the back-end application runs on a mainframe or in a cloud.

Rather, the nightmare scenario in his industry wouldn't be nimble startups but large overseas businesses from places like China or India that have the money to come in and take on the big players in the U.S. But even if they came in with modern architectures built from the ground up, he said, the extensive business rules that those companies would need to develop to play in the U.S. market would present a high bar.

For other markets, however, the bar may be much lower.

While the mainframe's viability may be open to debate in some quarters, the point, Wander said, is that your future competitiveness will depend on transforming most of the data center architecture you have in the next decade. "The folks who don't unclutter by the time we get to 2020 will have a tremendous mess," he said.