In a post on its blog Thursday, Netflix (NFLX) said it had migrated the last of the data stored in its own facilities to Amazon Web Services—a shift that Netflix credited with its fast speeds and eight-fold membership growth since it began the transition seven years ago.
But Netflix also acknowledged that its reliance on Amazon (AMZN) can come with problems, as illustrated back in September when a glitch in Amazon’s cloud database caused a temporary Netflix streaming outage for some customers. “We have hit some inevitable rough patches in the cloud,” Netflix wrote in its post, referencing an outage on Christmas 2012.
Still, Netflix said that such snafus have actually become more rare since it moved to cloud computing, and that it is approaching its goal of being up and running 99.99% of the time—meaning that its service would only be down for less than 53 minutes per year.
After all, before it was on the cloud, Netflix had “a number of outages” in its own data centers—including one in 2008 which is what prompted its “journey to the cloud” in the first place, the company explained on the blog: In that incident, “we experienced a major database corruption and for three days could not ship DVDs to our members.”
“Failures are unavoidable in any large scale distributed system, including a cloud-based one,” Netflix’s post continued. But by using various techniques in building and testing its software, “it is possible to survive failures in the cloud infrastructure and within our own systems without impacting the member experience.”