Microsoft has announced the general availability release of Windows Azure Media Services, the platform-as-a-service cloud that some broadcasters had used to stream London Olympics coverage.
Microsoft is hoping to attract the next big on-demand video outfits--Hulu or Netflix's successors--with the service, in addition to enterprise customers, network operators, broadcasters, ISVs, and system integrators.
Microsoft Developer Division Corporate Vice President Scott Guthrie announced the general availability on his blog, pointing to a new media services developer centre and enterprise SLAs put in place for the launch.
The service can stream video to HMTL5, Flash, Siverlight, Windows 8, iPad, iPhone, Android, Xbox, Windows Phone, and embedded devices for IPTV and Connected TV. Microsoft is also providing SDKs for developers to build custom video clients for their applications, which allows developers to customise playback.
The service will also come with the features that broadcasters used during the London Olympics: the ability to arrange live and on-demand content into logical workflows via Channels, and to ingest live video feeds, which are compressed and passed back to broadcasters to then publish to the Azure platform.
Microsoft is also offering customers a number of tools to help upload raw video, encode and protect it with DRM, and deliver it to different devices.
Developers can use REST APIs or .NET and Java SDKs to control these workflows on Azure Media Services, said Guthrie. DRM technologies that it supports include Microsoft PlayReady and Apple AES content protection formats.
At launch last April, Microsoft also pointed to a number of third party solutions that would form its base platform, including high-speed transfer services from Aspera; content encoding from Digital Rapids, ATEME, and Dolby; content protection from BuyDRM and Civolution; and video-on-demand streaming from Wowza Media Services.
Answering price questions on the blog, Guthrie said that encoding starts at a flat rate of $1.99/GB, discounted for larger volumes. There is no extra cost for virtual machines or compute power if encoding one job at a time, making it suited to small jobs, however, for high capacity jobs, encoding in parallel or with an SLA on time will incur extra costs.