Franklin & Marshall College

Improvements in performance, reliability and flexibility while saving energy

The Data Center is where the all of the computer network electronics integrate together with the large-capacity computers called servers to provide the mechanism for all of our data systems to function and store personalized information.  The room has dedicated power and cooling systems that provide a stable environment for upwards to seventy-five separate servers, many of which have been added in recent years as the demand for more on-line services has increased.  These servers require a controlled environment and power supply.  All "mission-critical" services (financial system, student records, email, eDisk, admission, advancement, learning management system) have redundant servers to ensure reliability.  Unfortunately most of these servers operate at between 10 and 20% capacity, so there are great inefficiencies.

In 2007, we researched and then implemented a more modern way to provide these critical services.  The new method is called server virtualization and the servers themselves are referred to as virtual machines or VM.  The virtual machine allows you to combine separate servers to improve performance in the following ways:

• Resources can be maximized and duplicated so that testing upgrades and new features are now in the background (improved "up-time")
• Space and power consumption are saved as separate servers are retired
• Services are now tied into our backup systems and are replicated off-site via a secure connection to Reading, PA where we have two servers and a network storage system
• All of this can be controlled remotely by our systems managers at any time

To date, we have retired almost all of our physical servers and have over one hundred virtual servers up and running, most of which are in production mode.  We are in a much more agile position to provide services to the college community now.  In addition, the off-site data center provides a safe mechanism to access institutional data if there were ever a situation where the on-campus data center was unavailable for an extended period.