Microsoft's Outlook email and productivity services stopped working for millions of users worldwide starting Thursday, with service restored by Saturday. The outage resulted from a planned increase in computing capacity that went wrong.
On Friday, the Federal Aviation Administration ordered all flights grounded after aged data-handling equipment failed. The FAA lifted the order Friday night, but hundreds of thousands of passengers were stranded over the weekend. The grounding was triggered by new cables installed during an upgrade attempt.
Both incidents stemmed from attempted upgrades that malfunctioned. The outages underscore the complexity of multilayered hardware and software systems. Changes made to improve performance can destabilize the delicate balance these systems require and trigger crashes.
The report notes that widespread outages remain uncommon, though system failures happen frequently. Two years ago, a routine software update by security company CrowdStrike crashed millions of computers running Microsoft Windows, grounding airline flights, halting subways, disrupting 911 operations, closing stores, and forcing hospitals to cancel surgeries. Restoring all affected machines took days.