Amazon has not yet said exactly what caused one of its cloud storage services to go down for several hours Tuesday, but businesses who were affected can learn from the headache.
The Seattle-based tech giant dominates the cloud services market, far outpacing its next three competitors.
So when an issue in a cluster of data centers in Northern Virginia took down one of the company's most popular storage services, the outage rippled across the internet.
The problem may appear to be online, but the geography is where the lesson lies, said Yevgeniy Sverdlik, the editor-in-chief of Data Center Knowledge.
Several companies who use Amazon Web Services were not affected by the outage. Sverdlik said those companies likely built their systems with redundancies: Even if they have information stored in Virginia, they might have copies in Oregon or Europe.
The takeaway: "Split your infrastructure among multiple cloud regions," Sverdlik said "If one region goes down, your application can stay up using a different region."
Businesses can do that with one provider, like AWS,. They can also try to use multiple providers, but Sverdlick said that's more complicated.
The outage affected Amazon's Simple Storage Service, known as S3. An S3 outage doesn't just affect businesses and websites who store information there because its is also the base for other AWS products.
"Lots of the companies that use any of those services can be service providers themselves," Sverdlik said. "So each of those companies may have hundreds of customers, and their customers are also affected, so the ripple effect is huge."
Sverdlik said that the outage was estimated to have affected thousands of sites and resulted in millions of dollars in losses.
