Alaska Air Has Air Max and Tivoli on the Job

Eye On: Alaska Air

In early 1998, Alaska Air Group Inc. (www.alaskaair.com) went live with a revenue management, forecasting and optimization system called Air Max from Sabre Inc. (www.sabre.com). Air Max was implemented to increase overall operations efficiency and flight yield management by ensuring that departing flights met capacity levels.

The Air Max system is a high-tech solution complete with predictive analysis and native business intelligence capabilities, but it’s only part of the story. Alaska Air’s central goal was to increase operating efficiency and overall revenue. As an integral part of this process, the airline uses the Tivoli Workload Scheduler from Tivoli Systems Inc. (www.tivoli.com) as a down-in-the-trenches tool to automate the vital job processes that ensure the smooth daily operation of the airline itself.

"This Air Max system takes a nightly data feed from our reservation system that’s transmitted to us every evening, and Air Max does its magical thing throughout the evening as far as forecasts go," says Terry Alexander, revenue manager with Alaska Airlines. "But the Tivoli Workload Scheduler is the wonderful tool that schedules all of those jobs at night so that we don’t have to intervene."

The Sabre system gathers data from external reservation and equipment tracking systems, and then forwards a consolidated "origin-to-destination" report based on the accumulated data to Alexander. The details of the report allow Alexander to notify reservation agents to fill seats at specified rates or turn available seats over to a web site for a last minute fire sale.

Because of the distributed nature of Alaska Air’s operation, the Air Max system must direct passenger status data across multiple networks, as well as collect all data at a central point to run simulations and deliver reports at set intervals.

A key component of the system, then, is an automated job scheduler that ensures that data gets processed and delivered to designated servers at designated times, without fail.

"Every night we get advance booking information for the next 329 days, and we get that from the Sabre system," says Rick Pykkonen, senior systems analyst with Alaska Air. "That info is currently coming to us through our mainframe system, and we pass that info down, and then we have a batch cycle that actually starts at 8 at night and finishes at 5 in the morning."

This is where the Tivoli Workload Scheduler comes in. While the Air Max system’s predictive analysis and business intelligence capabilities comprise one half of the equation, Alaska Air also needs a way to sequentially automate the batch jobs and processes that must be executed on a nightly basis to ensure the smooth operation of the airline.

"All of the batch cycles are scheduled through Maestro [the former name of the Tivoli Workload Scheduler]," Pykkonen explains. "Any one job schedule could have hundreds of jobs, so we have a real intense batch cycle to make sure that everything runs in sequence."

Pykkonen says that about 400 jobs are scheduled every night, many of them interdependent. The Tivoli Workload Scheduler’s graphical interface represents an improvement over Alaska Air’s prior mainframe solution.

"We were pretty much a mainframe shop for years and years and years. With Maestro, what people are used to is that it has a graphical interface so production control can actually be watching the screen, and if something doesn’t work right they’ll know right away," Pykkonen comments.

According to Alexander, the Tivoli Workload Scheduler ensures a level of efficient automation and it integrates natively with both the Air Max system and Alaska Air’s corporate job-scheduling solution, which is also based on the Tivoli Workload Scheduler. "If we didn’t have Maestro we’d have a body sitting in front of that machine all night long," Alexander explains. "Maestro is a wonderful tool because [Sabre] builds [Air Max] with it as the driving scheduler. And Alaska airlines, in a separate search, decided on Maestro as its corporate scheduler."