Following a string of highly visible tech outages, from Microsoft to the Federal Aviation Administration (FAA), many people are asking the same question. Kim Nash at The Wall Street Journal put it simply: how do we make sense of all this?
It’s tempting to assume a coordinated cyberattack. So far, the evidence points elsewhere: increased attention on ordinary outages, alongside a growing volume of technical changes as organizations work to patch vulnerabilities that artificial intelligence (AI) is helping them discover.
TL;DR – The recent tech outages don’t appear connected and none has been tied to a cyberattack so far.
Key takeaways:
It started on November 12, when Microsoft 365 went down. Outlook, Teams and other services were disrupted for much of the business day and into the evening. Microsoft traced the problem to an issue from an earlier network capacity upgrade and said a cyberattack was not to blame.
The next day, the FAA halted departing flights across the country, only the third nationwide ground stop in U.S. history. Thousands of flights were canceled or delayed. The FAA blamed an equipment issue during an upgrade to the telecommunications network that connects air traffic control facilities. Over the same few days, a widely used human resources (HR) and payroll platform, several major U.S. banks and a major airline also had outages. In each case, the organizations involved pointed to technical problems.
While outages are common across every industry, they’re rarely noticed. However, the recent string of outages, ranging from a few hours of downtime on a highly visible system to nationwide disruptions, have drawn national media attention. These stories have led to widspread speculation, which can be a natural response. When several major services go down within days of each other, it’s easy to assume there’s a connection, even when there isn’t one.
That doesn’t make these tech outages trivial. Many businesses depend on the same handful of platforms, so a disruption at one provider can reach organizations far from the source.
AI has made vulnerability discovery trivial for anyone willing to spend the tokens. In May, one major security vendor fixed 75 flaws, compared with its usual five per month.
Some vulnerabilities found with AI can be exploited by attackers and others are addressed before they pose a real risk. Either way, organizations are facing more remediation work than they are accustomed to and much of that work ends in a patch.
Every patch is a change to a production system and every change carries some risk of causing an unforeseen issue. That risk grows with the number of changes and the number of underlying systems involved. When those systems support other businesses, even small issues can have a much wider impact.
Other kinds of change carry the same risk. Microsoft traced its outage to an earlier network upgrade and the FAA blamed an equipment issue during a network upgrade. Neither has been described as a security patch, but both show how routine changes to large, interconnected systems can have outsized effects.
Many organizations freeze changes before the holidays to protect peak traffic. That can push teams to fit more changes than usual into the weeks before the freeze. Every organization handles its change windows differently, but this timing could be adding to the problem.
In GRIT’s view, the number of vulnerabilities waiting to be discovered is likely finite. AI is surfacing years of technical debt in a short period. Once major organizations work through it, the pace of change and the disruptions that can come with it, may return to relative normalcy.
The recent tech outages are a reminder of how much businesses depend on a small number of shared platforms, especially as vulnerability discovery accelerates and change becomes more frequent.
Managing vulnerabilities and minimizing disruption takes more than keeping up with patches. Organizations need a clear understanding of their exposure, a reliable testing process and a plan for responding when a critical system goes down.
Talk with our team about strengthening your vulnerability management and incident response programs.
None has been tied to a cyberattack so far. Microsoft said its outage was not a cyberattack and other affected organizations have pointed to technical problems.
Yes. Every patch changes a production system and any change can cause unexpected issues in complex, connected environments.
AI tools are helping vendors and researchers find far more vulnerabilities and many of them need a fix.
A structured patch management process with testing, scheduling and rollback plans reduces the chance of disruption.
Share what is known, what is unknown and what steps are underway. Early, honest updates maintain trust with customers and leadership.