New White Paper Released on The Performance Of Adaptive Streaming Over HTTP

Cisco and The Georgia Institute of Technology just released a technical white paper on adaptive streaming here. It has some very in-depth performance analysis and comparison between Smooth Streaming, Netflix player and Adobe’s OSMF. While there are some interesting findings about all three platforms, I think it clearly shows that Smooth Streaming is far more superior than Adobe’s OSMF’s default implementation in those tests.

Some of the interesting quotes from the article:

The Smooth Streaming player is quite effective under unrestricted avail-bw as well as under persistent avail-bw variations. It quickly converges to the highest sustainable bitrate, while it accumulates at the same time a large playback buffer requesting new fragments (sequentially) at the highest possible bitrate. This player is rather conservative in its bitrate switching decisions.

The Netflix player is similar to Smooth Streaming (they both use Silverlight for the media representation). However, we observed that the former showed some important differences in its rate-adaptation behavior, becoming more aggressive than the latter and aiming to provide the highest possible video quality, even at the expense of additional bitrate changes.

To summarize, we have observed that the OSMF player fails to converge to a sustainable bitrate especially when the avail-bw is smaller than or very close to the highest available bitrate of the media. Instead, it usually oscillates between the lowest and highest bitrates…. We do not describe here the rest of the experiments we performed with it, because they simply confirm that the default rate-adaptation algorithm deployed in the OSMF player does not function properly under our test scenarios.