When the cloud rains: Azure outage soaks industries worldwide
A major outage hit Microsoft’s Azure cloud service Wednesday, impacting email and other Office 365 apps used by companies and governments worldwide. A post from the company said users may experience “latencies, timeouts and errors.”
Alaska Airlines said the outage had disrupted key systems. In the United Kingdom, London’s Heathrow Airport said it was experiencing technical problems because of the Azure outage.
Azure’s troubles followed an outage last week at Amazon Web Services, the world’s largest cloud computing platform, which disrupted popular apps and websites. Azure is the second-largest cloud computing provider.
What is Microsoft Azure?
For those not in the tech space, that might be the first question.
“Cloud means this is the infrastructure,” Saurabh Bagchi, a computer engineering professor at Purdue University, told Straight Arrow News. “There’s the computer, the network, the storage, the memory on which many of the applications that we depend on, both at work and at home and for fun, are run on.”
It’s essentially a way for anyone who wants to be able to do work on the internet to have a platform to do so.
“You don’t have the money to go off and just set up a data center of your own, as most of us don’t have that kind of money,” Bagchi said. “Therefore, you go to these cloud providers who are going to provide you with all of this computing equipment on which you can run your application.”
Among the services impacted by this outage were Office 365, Minecraft and X-Box Live.
“All of these cloud systems are like a giant playground where pieces of tech, software, (and) hardware connect together, and it provides a place where companies like Costco, Starbucks, Alaska Airlines can all operate without them having to manage it,” Neil Johnson, professor of physics and complex networks at George Washington University, told SAN.
Impact of an outage
Those three companies Johnson mentioned were all impacted by the Azure outage.
“For our guests who are unable to check-in online due to the Microsoft Azure outage, please see an agent at the airport for a boarding pass, and allow for some extra time in the lobby,” Alaska Airlines posted on X.
Because the outage impacts numerous companies, Johnson and his fellow researchers put out a paper last year detailing “super failures.”
“It’s like a shock, which is more than the sum of the parts,” Johnson said. “It’s more than just Costco’s failure, Starbucks’ failure, Alaska Airlines’ failure. It combines together. It is stronger, so it affects more of the business space.”
These outages can also impact a wide variety of people for different reasons.
“If you are a video game connoisseur, you cannot play your video game on Roblox,” Bagchi said. “If you are a professional who relies on Zoom meetings, you are no longer able to do your Zoom meetings because all of this infrastructure runs on the cloud. If you want to go and transact some business on the bank, you’re not able to do that because the banking services rely on this.”
Data from IT Desk shows 96% of enterprises use a form of cloud service while 92% operate in hybrid or multi-cloud environments.
In a survey from Parametrix, 31% of U.S.-based corporate decision makers said eight hours of cloud downtime during business hours would be “catastrophic.”
“The idea is that these are going to be very well managed, so they’re never going to go down, so anybody on the internet can access your application any time of the day or night,” Bagchi said. “That idea is reached maybe 99.69% of the time. But that remaining very small probability of it failing is the one that makes this worldwide news.”
Cloud system protections
The top 10 companies in the country by market value all use some form of cloud service, and most have a partnership with AWS or Azure.
“Why would any of those companies want to have to maintain a cloud?” Johnson said. “I mean, they’ve got enough things to worry about, presumably, that they just want to use the service. ‘I want a space where I can put all my things, my databases, my payrolls, my kind of client bases.’”
With that comes the risks of outages, despite precautions taken by providers like Microsoft and Amazon. Their services are vulnerable to both error and attack.
Most outages, however, are not malicious, experts say.
“You leave milk outside for three or four days, nothing’s going wrong with it,” Johnson said. “Suddenly, there’s a failure of the milk. A big curd forms. It’s not like one of those pieces of milk was to blame. It was an organic thing.”
To minimize disruptions, cloud providers operate data centers in different regions, so that if one goes down, another can take over.
“One availability zone may fail because of a hurricane in one part of the country, but then the same application is running on another availability zone in a completely different part of the country,” Bagchi said.
Bagchi said the issues most often come from configuration problems.
“If you have some computing stuff of any reasonable level of complexity, then they will have many, many different kinds of configuration knobs,” Bagchi said.
“When you have some of these knobs that don’t get set quite right, then that has become a leading cause of outages,” he continued.
While outages aren’t common and providers do everything they can to prevent them, Johnson said this won’t be the last.
“This is an organic thing,” Johnson said. “And we can expect more of these and stronger and sooner.”
The post When the cloud rains: Azure outage soaks industries worldwide appeared first on Straight Arrow News.
Ella Rae Greene, Editor In Chief
