Amazon’s Andy Jassy, left, interviews Netflix CEO Reed Hastings at an event in Las Vegas.

Last month, founder Jeff Bezos and Netflix CEO Reed Hastings separately took the stage at the Re:Invent conference in Las Vegas and touted the significance of Netflix as a proud customer of Amazon Web Services.

Both suggested that the alliance symbolized the importance of how two companies — which compete head-to-head in streaming of movies and TV shows — can work together in delivering a cloud-based service.

“We may compete on Prime Instant Video, but we bust our butts every day for Netflix on the AWS side,” Bezos said. Hastings earlier touted Netflix as a symbol that it is “safe to be on AWS.”

So, when Netflix sputtered on Christmas Eve as a result of yet another Amazon Web Services outage at the company’s Northern Virginia data center, we had this simple question: Why did Amazon instant video, which also runs on the back of AWS, keep going?

It’s unclear to us what caused Netflix to sputter, while Amazon’s video service kept humming. In a series of reports on its health dashboard, AWS engineers noted that the outage was tied to errors in Amazon’s Elastic Load Balancing API calls.

A GeekWire reader pointed out the irony of watching a video on Prime Instant Video at the same time that the outage was occurring at Netflix. GigaOm also noted that Prime Instant Video appeared to be working, with Barb Darrow writing that outages of these kinds “can only help AWS rivals in the OpenStack community.”

Netflix, at least as of last month, didn’t appear to be going anywhere. Hastings indicated that he wants the streaming video service to become the biggest business in the world that relies 100 percent on AWS.

With the latest outages, will Hastings ever change his tune? The Netflix boss did just step down from the board of Microsoft, so not sure he’ll be heading over to Windows Azure anytime soon. But these outages do hurt Netflix.

As another GeekWire reader noted this week:

“Still out. Stupid. Seriously thinking of dropping NetFlix over this. Don’t care who their vendor is: they provide the service, they bill me. If they can’t deliver then I’ll drop ’em.”

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline


  • Guest

    Resign, Reed. I enjoy the circus as much as the next man, but it’s over, and we customers are heading for the exits.

  • Waaa

    To me all this overreaction really shows what a needed service Netflix is providing. If people are so frustrated in Netflix going down for a few hours they’re threatening to cancel (which they won’t) really means Netflix has become one of the necessities for many.

  • Forrest Corbett

    In other news, Amazon apologizes to Netflix customers for the outage, offers thirty-days free Prime Instant Video. ;)

  • Guest

    Let’s call it for what it is: sabotage. Amazon took Netflix down on Christmas Eve and Christmas day purposefully to kill NetFlix. Anyone that believes otherwise also believes in Unicorns and Fairies.

    Netflix were idiots to trust a competitor with their uptime. When Amazon moved into streaming video they should have figured what would be what.

    Much as I hate to suggest it but maybe they should’ve moved to Azure. That would seem the only cloud play not in competition with Netflix.

    Bezos is the next Bill Gates not only in terms of being a genius, but an evil genius. Edison, Gates, Bezos: they all made our lives better and were also evil.

    • Guest

      Let’s call you for what you are: moronic.

      Yes, I’m sure Amazon staged an outage that got widespread media play, risked its reputation and potential lawsuits, and affected hundreds of its corporate customers in order to knock Netflix, whom they showcased as a model client the month before at their developer conference.


  • kegill

    Cloud services have many characteristics of public utilities, two in particular: high fixed costs and very very low marginal costs. There isn’t the same social cost of duplication of fixed assets that you get with cable, electricity or phone (multiple lines running down a street for multiple providers) but there is a very large cost associated with changing service providers (lock in). In other words, the cloud services market does not have the characteristics that are inherent with truly competitive markets.

    I cannot reconcile the fact that Netflix was down (for me both TV and laptop) while Prime worked fine. I think both companies are seriously remiss for not speaking out quickly. And Netflix, for certain, should have had an overlay on the website advising of the outage within an hour — even more appalling that there was no alert (nor a change in the error message when trying to stream on the desktop) given that the service was down for almost 24 hours.

  • davidgeller

    I’m anxious to hear the real details – from either Amazon or Netflix. The simple fact remains that while it’s incredibly easy to get started with AWS and host all your company’s services on their platform, it remains challenging to design and, eventually, configure everything to successfully survive a single or multi-zone failure. There have been many high-profile sites and platforms that have failed to survive such failures on AWS simply because they didn’t take advantage of all the capabilities AWS has to offer. So, it’s entirely possible that Netflix fell into this trap as well. Yes, AWS had a failure in its Northeastern data center (NJ?). That doesn’t, necessarily, suggest that Netflix couldn’t have survived the outage if it has configured things differently. With regard to Amazon’s own video service surviving while Netflix’s didn’t – that could simply be an audience size issue. Netflix is probably much bigger and serves much more video.

Job Listings on GeekWork