Germany taz
Worldwide Microsoft disruption: An old program error caused the outage
Recently, Microsoft users around the world have been experiencing problems. It has now been revealed that the error responsible for this has been dormant in the system since July. more...
Deutsch (original) ·Auto-translated to English
dpa | A program error that has been slumbering undetected in the system since July triggered the global outage of Microsoft services such as Outlook and Teams last week. As the US software company announced, three factors together led to a chain reaction: At the peak of the disruption, more than 80 percent of the computers responsible stopped responding on Thursday evening.
According to Microsoft, the trigger was a software change to a central digital entrance door for Microsoft 365 data traffic, which had already been installed in July. The system ran smoothly for more than three months. The error was not noticed in the previous tests in the laboratory because older technology was primarily used there - however, the problem mainly occurred on newer, more powerful servers.
The error broke out last Thursday during peak usage time in the USA due to two additional loads: Shortly before, the video conferencing program Teams had been switched to the same main infrastructure as planned, which increased data traffic. At the same time, almost 14 percent of the server capacities were temporarily unavailable due to another technical problem.
When the first servers failed under the high load, Microsoft's automatic control redirected the user requests to the remaining computers - and thus triggered the same error there too. In less than two hours, this domino effect affected almost the entire system. Because digital protective barriers between the individual regions of the world failed, the outage spread around the globe.
While services such as Teams or the online storage OneDrive could be redirected to alternative routes relatively quickly, this was not possible with the heavily loaded Outlook email system. The fix took until Saturday because freshly restarted servers immediately collapsed again under the flood of waiting requests. Microsoft eventually had to develop a special repair update and gradually restart the servers in groups.
The effects of the disruption were felt in more than 70 countries: Numerous companies, courts, schools and authorities were cut off from their emails, calendars and online documents for hours - corresponding reports were also received by the Federal Office for Information Security. In many places, the healthcare system was hit particularly hard: in Great Britain and Canada, for example, hospitals and doctor's practices had to switch to paper files and postpone individual appointments.
In many offices, employees were forced to resort to private messengers such as WhatsApp or the telephone; Risk analysts estimate the global economic productivity loss to be several billion dollars.
After the serious Microsoft failure, there was renewed criticism of the strong dependence of many companies and administrations on a single US company. IT experts, data protection advocates and European competitors warned of the so-called cluster risk if all office communication depends on a single cloud platform.
Schleswig-Holstein's Digital Minister and State Chancellor Dirk Schrödter (CDU) called on Deutschlandfunk for more digital resilience in view of the "deep dependency": Public IT systems must be treated like critical infrastructure and converted to applications with open programming code (open source) - as has already been done in the state administration in Kiel.