Home > Strategic Technology Planning >> Critical Thinking

Features

Strategic Technology Planning >> Critical Thinking

8/29/2006

Virtual Tools

Servers always present a double-edged problem for colleges and universities. On one hand, in order to offer missioncritical functions such as content management, ERP, and more, schools must invest in literally dozens of servers in order to do the job right. On the other hand, multiple server investments can be hefty, stretching already thin IT budgets even tighter. At the University of Wisconsin Medical Foundation, part of the University of Wisconsin-Madison, technologists had faced this dilemma for years. Finally, in January 2006, Network Analyst Troy Frank and his colleagues came up with a solution to the problem once and for all: virtualization.

TECH INITIATIVE TIP

Don’t just think "overhaul." Even a small improvement can make a huge difference, if it’s made within a critical application.

Virtualization is the process of presenting a logical grouping of computing resources so that they can be accessed in ways that provide benefits over the original configuration. In the realm of servers, this means partitioning one physical server into a number of smaller “virtual” ones, making the most of the resources at hand.

At the University of Wisconsin, the Medical Foundation turned to this strategy earlier this year. Already, with the help of Virtual Infrastructure from VMware, Frank says the foundation has consolidated 30 servers into four physical machines—a savings of both money and floor space in the server room.

“By virtualizing, we’re no longer plopping down servers indiscriminately,” he says. “Everything is part of a bigger plan, and it all fits on a perfectly manageable number of machines.”

Frank says that he and his colleagues haven’t sat down to figure out what kind of money they’re saving on hardware costs by moving to virtualization, but he notes that each of the four VMware servers they purchased cost $5,000 apiece. He estimates the foundation probably has saved on utility costs as well, since it is powering and cooling fewer servers than ever before. And he points out that another area where the foundation likely has saved money is staffing: In the past, a group of five or six technicians managed the server farm; today the machines are handled by one full-time employee and one part-timer who fills in when the department needs him.

Down the road, Frank predicts the foundation will move most of its remaining 150 servers into the virtual architecture. One potential problem is bandwidth; depending on the preexisting workload of a particular server, not every piece of hardware is a good candidate for virtualization. Because every virtual server on one machine is sharing resources such as memory and disk space, Frank and his colleagues say they must be careful not to overburden any one entity. Their golden rule: Don’t virtualize heavy-duty servers that need dual- or quad-processing power unless they have 4GB of memory to access at any time.

“We can’t just virtualize everything,” says Frank. “We know that in order to get the best performance out of the technology we have, we need to be smart about how we consolidate functions into one.”

Locking Down a Network

At Fordham University, technologists recently turned to technology to shore up the school’s network defense—a critical objective for administrators there. In 2001, a staff shortage in the Information Technology Services (ITS) department had made it difficult to monitor the university’s systems, particularly during the overnight hours. Then, when network outages did occur, ITS resources could not provide immediate notification of the problem. Instead, department officials had to wait until morning to identify and correct any issues. This delayed and reactive approach to addressing IT problems simply was not good enough, administrators decided; Fordham needed a solution to augment capabilities and deliver more proactive, efficient service.



Recommended Reading