The 1990 AT&T Network Outage

On the fragility of our technological systems

On January 15th, 1990, sixty thousand people lost their telephone service and seventy million phone calls went uncompleted. Nearly fifty percent of all calls placed through AT&T failed to go through. For nine hours, AT&T’s long-distance telephone switching system crashed.

Why did America’s telephone network come crashing down?

Because of a software update. In early December, AT&T technicians upgraded the system to improve processing speeds for messages. While the code had been carefully reviewed, a one-line bug slipped through. So, on that fateful January day, when the New York City switch team performed a routine self-test, the system went haywire. The giant map of the United States hanging in the company’s network operations center in New Jersey lit up with error messages.

The network connecting our nation crashed.

Today, flights are grounded, court sessions are being postponed, websites are down, television networks can’t broadcast shows, and surgeries have been rescheduled.

Why did our world come to a standstill today?

Because of a software update. This update — designed to protect against hacking — was sent out by a largely-unknown cybersecurity company, CrowdStrike. But when the update hit Microsoft Windows computers, it caused them to crash. There was “a defect … in a single content update,” and because so many companies and organizations use CrowdStrike, the flawed update reverberated around the world. Many consider today’s event the largest IT outage in history. More than 70% of the world’s desktop computers run on Microsoft Windows, and many of those computers are currently experiencing the repercussions of the faulty code.

Commenting on today’s events, professor Ciaran Martin told The New York Times: “This is a very, very uncomfortable illustration of the fragility of the world’s core internet infrastructure.”

In 1990, computer scientist Peter Neumann offered a similar provocation in response to the AT&T outage: “When you have one extremely complex system, there can be very simple hidden failure modes that nobody has any idea about.”

There is a delicate interdependence supporting our global tech ecosystem. Networks run on technologies, and if those technologies fail, those networks fail. Even the smallest error in a line of code can bring down a national telephone system or the global economy.

That nine-hour outage in 1990 cost AT&T $60 million in unconnected calls, and that’s not counting the millions of dollars lost by individual businesses that relied on the telephone. The economic effects of today’s outage are yet to be calculated, but the millions — or even billions — of dollars likely lost by businesses today underscore how our modern economy runs on an interconnected, often-fragile technological infrastructure.