Systemic Vulnerabilities In The Digital Age
In case you haven’t seen the news today, the morning started with United Airlines saying that they experienced a “system-wide computer problem”. United later said that an issue with a network router “degraded network connectivity for various applications, causing this morning’s operational disruption”. The disruption meant that no United flights took off from about 8:00 AM Eastern time until 9:47 AM. This affected about 800 flights. If you assume an average of 125 people on a flight, that means that this problem affected at least 100,000 people. People trying to make connecting flights were impacted more than people on direct flights and people on tight schedules may have missed whatever they were travelling for.
Next, the New York Stock Exchange halted all trading at 11:32 AM Eastern Time and cancelled all pending trades. The exchange said that was the result of an internal technical issue – a network connectivity issue. They planned on resuming trading at 3:10 PM Eastern Time.
Finally, The Wall Street Journal’s web site home page started generating an error at about 11:45 AM. It came back up around noon.
What the root cause of these outages are or if they are related is unclear. No one is saying much. The government is saying this was not the work of hackers, but I am not sure that they would say so, at this point, even if they thought it was.
What this does point to is that we are pretty dependent on technology these days. If you go into a store and the power goes off, they close the store because they have no way to take your money. Same thing if their computers crash.
While that is inconvenient, it is a bigger issue if the organization that suffers an outage is part of the critical infrastructure (water, power, police, healthcare, etc.)
Organizations like the NYSE have spent tens of millions of dollars – maybe hundreds of millions – making sure that their systems stay up when they are supposed to. But they don’t always stay up.
It is unlikely that any traders will walk away from the NYSE. Some travelers might leave United over the outage, but likely not many. WSJ readers tend to be pretty loyal.
However, the cost of dealing with these outages is very high. Big commercial trades that would have been made on the NYSE were likely made on other exchanges. United probably had to refund passengers or accommodate some passengers on competitors planes. The WSJ probably lost the least, but likely did lose some advertising revenue.
These outages were all relatively short – less than an hour to a few hours. They also did not affect things critical to you and me on a daily basis (like when the Subway in New York or the Bart in San Francisco shuts down).
What if the outage was longer. What if a bank system failed and was down for several days or an airline operations system was down all day. Both have happened.
We have spoiled ourselves because generally, things in America are pretty reliable.
That does not mean that they are not fragile. And fragile is not a good thing because that means that small perturbations can break them. Potentially affecting you and me.