Amadeus expects this year to reach a major milestone, decommissioning the last of its mainframes. This is the culmination of a multi-year process to achieve “more flexibility and resiliency.”

Sabre isn’t there yet. Under new CEO Sean Menke, the company has embarked on a general tech upgrade program. “We will make investments in our IT infrastructure this year to modernize, drive efficiency in development and ongoing technology costs, further enhance the stability and security of our network and accelerate our shift to open source and cloud-based solutions,” said CFO Rick Simonson in February.

Improving speed-to-market, flexibility and cost efficiency are key goals, according to Sabre.

Amadeus and Sabre together provide reservations technology to airlines flying more than half of all passengers boarded globally, according to Travel Technology Research. They are the two largest players, with China’s TravelSky a distant third and the rest of the market highly fragmented. Amadeus and Sabre also lead the markets for airline departure control and inventory management systems, according to T2RL.

With many media outlets blaming system problems on antiquated technology, readers may hope that these upgrades will tackle the gremlins behind recent airline IT outages. Unfortunately, the premise is oversimplified. Say an airline’s systems go down and flights stop. It blows up on social media. Maybe the company declines to tell reporters exactly what happened. Journalists then quote industry observers who guess at the cause. Often not airline IT experts, these observers cannot resist assuming that old technology is involved, and that old tech is bad tech.

An article last week in The Wall Street Journal attributed problems to complexity and implicated “aging” reservations systems. But those systems did not cause the referenced outages, according to the article itself. The problem-causing systems cited by the Journal were a computer router, a power system, a piece of texting software and the Internet.

IBM mainframe
Image: IBM

When the media mention “antiquated technology,” they typically are referring to mainframes running the Transaction Processing Facility (TPF) operating system. When IBM invented these in the 1960s, some of the hardware looked like oversized reel-to-reel audio recorders. These systems and their successors are known for being secure, reliable and crazy fast. The hardware has evolved. The latest mainframes look more like Batman’s refrigerator, as one Bloomberg editor joked.

Originally created by IBM and Sabre, TPF software that runs on these mainframes is still a top choice for computing environments asked to handle the highest volumes of transactions by the largest number of users. Almost all of the biggest banks, insurers, retailers and airlines still use it, according to IBM. Global distribution systems were built on it.

Nowadays, IBM’s z/TPF mainframes also can run more modern software. Airline IT companies for years have been moving functions in steps either to that new software within the mainframes or to new hardware as well.

“There’s nothing wrong with TPF,” said one former Amadeus executive who preferred to be unnamed since he’s no longer with the company. “It’s a workhorse. Everyone knows it’s limited but I don’t blame TPF for outages. It’s stable.”

The Amadeus migration is more about commercial endeavors than IT reliability, he said. “It’s like a religious debate. There are reasons for and against.”

TPF has its flaws. Finding programmers can be a challenge. It’s rigid. Of late, the limitations have made it difficult to keep pace with airlines’ retailing ambitions. These downsides change the equation.

Sabre does not run everything on mainframes. Some components don’t need their processing power. The company has moved shopping, inventory, customer profiles, availability, ancillaries and other applications to open-systems environments.

It was a failure in the pricing system that hurt a handful of Sabre’s airline clients last fall, Sabre executives said last month. The company will spend $20 million this year to move that pricing system from a Hewlett Packard Enterprise-managed facility to an internal one, they revealed. Mainframes were not to blame there; nevertheless, Sabre is planning for life without them.

“We will continue to replace mainframe services with open systems services where it drives innovation and value for our customers,” Sabre senior vice president for delivery solutions Dolly Wagner-Wilkins said Tuesday. “The discussion at Sabre is not if but how. Over time, as the marketplace evolves, we expect the mainframe to be completely replaced with new services.”

Amadeus “has almost completed its move,” according to a February statement attributed to Amadeus IT Group VP for reservation, distribution and mid- and back-office Denis Lacroix. “The rationale here is that these mainframes are very heavy and centralized, and today’s world is no longer centralized.

“A number of years ago, Amadeus decided to decommission its mainframes and move all its core applications to a large network of Linux computers,” Lacroix explained. “Beyond Linux, we have a number of open source initiatives ongoing, and have built our own cloud platform called Amadeus Cloud Services. The open source model has had a massive impact on how the company works, giving us a host of benefits — from the collaborative work sharing ideas and experiences with other open-source companies, to being able to hire the right talent that can help us to stay ahead of competition.”

Additional info: For Sabre, the overall modernization effort is a driver behind recent layoffs, according to Simonson’s fourth-quarter presentation to investors.

No longer a provider of res systems per se, Travelport years ago moved certain services to open systems, including fares and shopping. Most of the new capabilities the company has built use open systems, but the mainframe remains “the system of record,” according to CEO Gordon Wilson. He said Travelport has not found anything better than z/TPF to do the “back-end system work” for its GDSs. “There are very few players who were TPF-based that are not still using the system of record, albeit having moved the more dynamic stuff into open systems,” Wilson said during a February interview.

Wilson also said Travelport is adding services to provide disaster recovery for Delta’s systems. “In the event of a systems issue or a fire, they have asked us to provide those services,” he said. “They are moving from another provider. We’re building another data center which is an investment we are making. We are going live with that in the summertime.”


Stats Show Outage Won’t Cause Corporate Payouts Under Delta Guarantee

Fare Combinability Issue Cools

Explainer: IATA’s New Distribution Capability

Sabre To Deliver AA’s Paid Seats Starting Feb. 2

Egencia Bags Incremental Advance On Ancillaries

Concur Adds Delta, Lufthansa Branded Fares, Favors GDS Connectivity