Microsoft has been experiencing a service disruption impacting most of their cloud (internet shared) services. This includes things like corporate Outlook Mail and Teams as well as their AI services such as Copilot. Apparently the cause was some kind of failure relating to an upgrade they performed- in other words, they aren’t being attacked, it is their own fault.
The article linked above touches on a few reasonable concerns regarding the downsides to centralizing these kinds of services amongst a handful of providers.
Excessive centralization
As recently as a decade ago the idea of one provider bringing down email for millions would have been a strange thought. Most large corporations back then had their own email services run internal to their own networks. Twenty years ago this kind of arrangement would have been a near certainty.
But starting around 2015 more and more companies began transitioning to using one of two or three providers for their corporate email services. Microsoft made this almost mandatory by essentially eliminating products and offerings that could be managed internally by corporations in favour of “cloud based” services such as Microsoft 365. And as we all know, Cloud really just means “some else’s hardware somewhere else”: for email and related office productivity services cloud-based became the solution of choice.
Now instead of there being hundreds or thousands of separate office productivity solutions out there in Corporate land there are practically only two or three. Microsoft (Microsoft 365) and Google (Google Workspace) are the main ones that come to mind, and I suspect that Outlook is by far the most dominant. This kind of oligopoly leads to a lack of resilience: a failure takes out millions instead of just a few thousand. It also leads to other problems like lack of innovation spurred due to lack of competition as well as a stagnant pricing model for the same reason.
Conclusion
Microsoft has had a severe degradation or outage across almost all of their service offerings that has lasted for days. They are slowly starting to recover now and normal services are likely restored for most users. I respect the work the technologists had to do to bring things back online with a live system, and hope they learned something to prevent future incidents like this.
The lack of a vibrant mix of service providers or self-managed (internal) services has left corporate office productivity offerings very vulnerable to this kind of outage. It is probably too late for most companies to reverse course and consider running such services internally or via a different provider, and the tradeoffs in terms of functionality and quality may not be worth it. But it is a reminder that doing what everyone else is doing isn’t always the best course of action.
