Erie Community College’s Joseph Stewart recognizes that downtime is no longer an option.

There's No Time For Downtime at Erie Community College

Erie Community College is working toward total availability of IT services.

Erie Community College is working toward total availability of IT services.

An e-mail outage several years ago convinced CIO Joseph Stewart that downtime for the systems and applications at Erie Community College had to be eliminated. The frustration and irritation among students, faculty and staff was just too profound for him to miss the message.

“It was five or six years ago, and e-mail was down for less than 24 hours,” Stewart says. “But even back then, the backlash made me realize that breaks in service just aren’t acceptable anymore. People demand almost 100 percent uptime, and we have to find a way to balance that need with our budget constraints.”

As part of a comprehensive effort toward the goal of providing “total availability” of computing resources for its users, Erie Community College (ECC) is purchasing and implementing power management and protection systems centered on APC Symmetra uninterruptible power supply (UPS) systems. The series of deployments will not only bolster 24x7 access to networks and applications, but also safeguard the integrity of critical data that can be lost or corrupted during outages.

Up and Running

ECC, which is part of the State University of New York system, serves more than 13,000 students on three main campuses and several more satellite sites in Buffalo and its surrounding suburbs. The college uses e-learning software extensively and maintains more than 60 computer labs equipped with recent model technologies — so keeping systems up and running is a priority.

“IT is the critical backbone to everything we do here,” says Bill Reuter, ECC’s chief administrative and financial officer. “It’s important to keep up with what students and employees are used to in other aspects of their lives, and to get students ready for the technology they’re going to encounter in the world. When any part of the system goes down, they act as though they can’t operate at all, as if pencil and paper were no longer an option — and in many ways, they aren’t.”

Managing continuity of operations and providing high availability to systems is essential. But with limited budgets, it’s often difficult to meet these requirements in higher-education settings, says Roberta Witty, a research vice president at Gartner. Buying and upgrading UPS and power management solutions is perhaps the most basic part of that strategy.

“That’s got to be standard operating procedure for anyone who’s managing the availability of IT assets,” she says. “A couple of minutes or even seconds without power in a data center is a big deal.”

To keep operations and education functions running smoothly for ECC’s 18,000 users, Stewart and his 16-person staff maintain a complex and growing IT infrastructure. An enterprise resource planning system supports web- and paper-based student registration, along with back-end financial, human resources and payroll applications.

A storage area network (SAN) delivers 100 terabytes of storage for both administrative and educational data. E-mail at ECC runs on Microsoft Exchange. The college uses Microsoft SharePoint and is implementing an intranet portal using SharePoint as a platform, says Stewart.

ECC’s campuses are connected over a Cisco Systems gigabit network backbone. Voice and data communications are unified through IP telephony, which also supports the school’s emergency alert and video surveillance systems.

No Downtime Tolerated

“If our systems go down, we’re finished — we cease to exist until they’re up again,” Stewart says. “The students are not very forgiving. Fortunately we have had very little downtime and expect less going forward.”

ECC already has backup generators and UPS units to protect its data centers from random utility interruptions as well as from the violent lightning, blizzards and ice storms that blow in from Lake Erie. The APC Symmetra LX 16 kilovolt-ampere unit installed in ECC’s north campus data center in April has twice the capacity of most systems.

Stewart has already installed a much larger Symmetra PX 80KVa unit for use in ECC’s new data center, which was recently created in a downtown building donated to the college by the county government.

Symmetra LX UPS systems are designed specifically for server rooms and data centers with limited space. All the Symmetra units, regardless of capacity, can be managed remotely over the network and can scale power and runtime in response to increased demand. Because the APC units can be quickly upgraded in the field, they reduce the need for UPS oversizing, a practice that leads to wasted capacity and adds substantially to the cost of power management solutions.

Knowledge Counts

The Symmetra technology was chosen using the thorough purchasing process ECC applies to all of its IT investments, Stewart says. First, he investigates any New York state procurement contracts that ECC might leverage to keep costs down. Then he reaches out to other SUNY CIOs to get the benefit of their experience with and knowledge of any products under consideration.

Meanwhile, the IT staff researches manufacturers’ reputations, technical literature and sources such as the blogs of fellow practitioners. Whenever possible, ECC arranges trials to test products in the college’s environment before buying them.

Over time, the college intends to replace the old UPS units in all campus data centers, server rooms and wiring closets with newer Symmetra LX and PX technology, Stewart says.

Besides the continuing issue of funding, another obstacle to the plan is an aging electrical infrastructure that was not built to support the amount of computing equipment installed in school facilities. As power requirements grow, the wiring in place will need to be augmented with entirely new feeds from the outside, Stewart says.

Once the fundamental electrical issues are solved, APC will handle installation of each UPS unit as part of a start-up service contract, says Joseph Lundin, ECC’s director of communications systems.

“They send out an APC engineer who sets up the unit, tests it and tells us when we can plug in our equipment,” Lundin says. “That means we minimize the downtime when we switch over to a very brief interval.”

Disaster Recovery Plan

Power management is just one fundamental aspect of ECC’s strategy of maintaining high availability and improving disaster recovery preparation, says Stewart. Moving into the new main data center on the city campus means that the old primary data center can be used as a mirrored DR site. Stewart is developing a new, detailed disaster recovery plan.

“In the past, we had all our eggs in one basket. We had backups. But if there was a major catastrophe, getting operational again in two or three days was probably not achievable,” Stewart says. “I feel pretty confident that within six months that recovery window will be much shorter.”

Maintaining continuous availability of services, rather than disaster recovery, has been the primary focus of ECC’s efforts, says Stewart.

The Symmetra UPS solutions, clustering and virtualization can all produce savings in power consumption and staff time — the latter resulting from both fewer outages and more centralized management of computer assets. But return on investment is not the primary driver.

“We think about ROI, but the main issue here is high availability,” Stewart says. “What was acceptable eight years ago — having periodic outages — is not acceptable now. Information technology has become another utility, like water or electricity. We look at the investment as a necessity.”

Total Availability Checklist

The different categories of technologies listed below can all contribute to the goal of providing 100 percent availability and ensuring that recovery from any disaster is quick and complete. to choose specific manufacturers and products, it’s best to follow the example of the Erie Community College It staff: rely on research, information from colleagues and field trials whenever you can get them.

Total availability should include the following:

  • Uninterruptible power supply units
  • Backup hardware
  • Backup software
  • Storage area networks
  • Replication and clustering technologies
  • Virtualization
  • Continuity of operations management software
<p>Mark Battrell</p>
May 19 2009

Sponsors