A software update by cybersecurity company CrowdStrike on July 19 crashed Microsoft Windows operating system around the world. Each of the major US airlines was forced to cease operations.
The experiences of American Airlines, Delta Air Lines and United Airlines as well as those of their customers could not be more different. FlightAware’s data shows that American had largely recovered by the evening and had only 51 mainline cancellations on the following day.
Delta, meanwhile, captured unwanted national headlines as network restoration dragged on over five miserable days in which the carrier canceled approximately 7,000 mainline and regional flights. It said those cancellations disrupted the travel of 1.3 million customers and cost the carrier approximately $500 million. United was in the middle of these two responses, taking three days to return to normal and canceling more than 1,400 flight.
Why Delta performed so poorly has become a source of conflict between the airline, CrowdStrike and Microsoft — a clash that appears to be spiraling toward litigation. The airline declined to comment on this story, but only gave a general explanation as to why American had much less trouble.
“One thing we’ve learned in the past is that you need to track your aircraft and crews in case of disruption. You should take action as soon as possible so that you don’t lose your ability to recover,” American CEO Robert Isom stated during an earnings call.
“We have built technology and we have done the right things to ensure that we take early precaution, early steps.”
These comments teased at the biggest sticking point that Delta has stated in relation to its protracted recovery. The airline said 60% of its mission-critical applications, including redundant backup systems, rely on Windows, and that during its recovery it had to physically reset 40,000 servers — a bigger lift than any other airline.
Ed Bastian, the CEO of Delta, explained that the airline’s long recovery process was slowed by the loss of an important crew tracking tool. This left the airline unable to track the location of its flight crews. Delta could not reset its operations without that information.
Microsoft accused Delta, as part of a bitter public exchange involving CrowdStrike and Microsoft, of trying to deflect its own blame for the long-lasting operation collapse.
Mark Cheffo – an attorney for Microsoft – wrote in a letter to Delta dated 6 August that “our preliminary review suggests Delta, unlike its rivals, has not modernized their IT infrastructure.”
Experts in airline operation and IT said that many factors could have affected why Delta recovered so badly from the CrowdStrike incident, while other airlines affected fared much more well.
They said that Delta’s heavy dependence on Windows applications could have played a part. Even random factors such as the people who were working that night could have affected how each carrier responded.
Daniel Stecher said that airlines might have had more employees or the right experience in the shift. IBS Software provides cloud-based software for airlines.
“That person will handle the disruption completely differently than someone who is new.”
Matt Cincera told CIO, a trade publication, in March 2023 that Kyndryl, and not Microsoft, was responsible for the airline’s tracking system. Delta, despite its hostile exchange with CrowdStrike still claimed that the cybersecurity company was responsible for the failure of the crew-tracking system because the operational disruption resulted in a large amount of incomplete data to be delivered to the system.
The dispute is indicative of the complexity found in airline IT systems, which combine mainframes from decades ago with modern, cloud-based apps.
“At the very least, in my experience, legacy airlines’ flight operating systems, such as crew-tracking and the maintenance-tracking system, are among the oldest and most restrictive systems,” said Bob Mann of RW Mann and Co. Bob Mann has worked on key IATA IT committees and was a consultant for RW Mann and Co.
Stecher said the mix of legacy systems and cloud-based ones common in airline operations centres creates IT silos that impede decision-making as well as response efficiency during disruptions. He said this is a problem that leads to suboptimal reliability in the industry.
“Aircraft have a lot redundancy.” This is why they are safe. This redundancy costs you money. But this redundancy was not implemented on the ground,” said Stecher. IBS, Stecher’s employer, sells a cloud based integrated operational platforms.
Airlines, he said in general, aren’t as willing to invest as much in IT as they should.
Mann said that it’s not easy for an air carrier to transition away completely from its legacy operating systems when it operates thousands of daily flights. He said that such changes must be made gradually over a long time.
“At this scale, you can’t afford a bet that might work,” Mann explained.
