Raising the Confidence LevelBy David F. Carr | Posted 2004-01-16 Email Print
WEBINAR: On-demand webcast
Next-Generation Applications Require the Power and Performance of Next-Generation Workstations REGISTER >
Poor on-time performance was crippling America West. Then the airline dumped its failing hardware.
Raising the Confidence Level
On one side of the room at the Strategic Operations Center (SOC) in Phoenix, dispatchers deal with last-minute changes to flight plans, creating the critical "paperwork" that America West aircraft need to be cleared for departure from airports around the U.S. Across the room, maintenance coordinators track progress on service and repairs, and crew coordinators reassign personnel when a flight is rescheduled or a pilot calls in sick.
In 2002, the SOC was relocated from the Phoenix Sky Harbor International Airport to a new $35 million building nearby that also houses simulators for America West's flight-training program. Getting it off what Beery describes as the airport's "rickety" network infrastructure was part of the effort to improve reliability. If the computer systems supporting this operation fail, planes and passengers sit idle across the country.
One problem with the airport was that it used a token-ring network in which computers are wired together in a circle, a topology that's been declining in popularity for years. One disadvantage is that the network can be disrupted by a break anywhere in the ring. Most organizations now use Ethernet networks in which each computer connects to a hub or switch.
In 2000, when a network card failed in the database server for flight planning, America West's systems personnel discovered they had no spare on hand and couldn't quickly get a token-ring card that fit the server's proprietary hardware interface.
From there, the problem snowballed. Even after network functions were restored, it took time to restart airline operations, given that many planes weren't where they needed to be. That meant the entire operations staff had to scramble to adjust flight, staffing and maintenance plans.
Today, the network is better protected. In May 2002, the SOC applications were switched from the servers at the airport to more-powerful Hewlett-Packard DL-380s in a nearby data center. Those servers are equipped with two network cards, each with an independent network connection, meaning that a server can still function if one card fails. Six weeks later, the operations center itself was relocated from the airport into a wing of the new flight-training center, just outside the airport. Now all workstations are connected to a 100-megabit Ethernet network, which is faster and more dependable.
On a tour of the SOC in September, Beery strolls the perimeter, pointing out the dual flat-panel display terminals at each desk that reveal detailed information at a glance. He spirals into the center of the room where SOC manager John A. Anderson, sitting behind a bank of three screens, monitors the activities of everyone in the room.
Anderson's workstation makes it easy for him keep the airline operating smoothly. He toggles between air-traffic control screens, weather maps, aircraft-maintenance status and performance metrics, navigating to each application through a Web framework based on Computer Associates' CleverPath portal server. Most important, he can rely on the system.
"If my dispatchers can't get the paperwork out of here, the planes can't go," Anderson says. Today, work is rarely interrupted by a system outage.
The new network also cuts the time it takes a dispatcher to recalculate a flight plan from 25 or 30 seconds to 5 or 6. "When we push that button to make the calculation, speed is extremely important because we may go through that several times" before producing a flight plan that's agreeable to the pilot and to air-traffic control, Anderson says.