"Our datacenter runs Oracle on IBM hardware, we could have switched to commodity hardware in a data center, but skipped that step by going to AWS," Netflix cloud architect Adrian Cockcroft told the consulting firm Cloudscaling in an . "There are three points on cost, one is that Oracle on IBM is very expensive, so AWS looks cheap in comparison, and we have flat-lined our datacenter capacity."
In addition, Netflix "could not have hired enough [system and database administrators] to build out our own data center this fast. We have added 4-5x as many systems in the cloud as the total we have in our data center over the last year," he said.
Finally, EC2's pay-as-you-go model means costs are elastic. "If you own a resource it sits around a long time waiting to be delivered and installed, and if you no longer want to use that type of resource you are still paying for it for three years."
Cockcroft's remarks add some color to Netflix's announcement in May that it The company had already been using the service for various customer-facing and internal applications, but decided to add "critical pieces" of its service, including member movie lists, a recommendation engine and film transcoding.
The move enabled Netflix to free up "scarce engineering resources from the undifferentiated heavy lifting of running its own infrastructure," it said at the time.
Netflix's skyrocketing customer count, which now stands at about 16 million, also made the job of running and expanding data centers too unpredictable, according to a presentation Cockcroft In addition, the company has been rapidly transitioning from a DVD-delivery outfit to a mainly streaming operation.
For actually streaming the movies to customers, Netflix contracts with companies like Akamai and Limelight, Cockcroft told Cloudscaling. It also runs systems for account sign-up, billing and other needs elsewhere, he added.
The IDG News Service