Steps to Ensuring Seamless Video Delivery OTT (Over the Top)

OTT Streaming Video Services

video delivery

How long will you wait for your chosen video to start playing? How long will you stare at a frozen video screen? How long will you suffer a herky-jerky video that cuts in and out?

And you pay a monthly service fee for this Over–the-Top (OTT) video service!!!???

Consumers of over-the-top video are not a patient bunch!

According to Conviva’s “How Consumers Judge Their Viewing Experience: The Business Implications of Sub-Par OTT Service,” nearly three-quarters of all OTT video viewers give up on OTT video viewing within the first four minutes if they experience “stream interruptions, inadequate picture fidelity and other poor streaming experience issues.”

OTT refers to services that are delivered directly from the content provider to the consumer over (the top of) their normal Internet connection, which could be cable, fixed or mobile. YouTube is probably the best known free OTT video service. Popular subscription services include NetFlix, Hulu, NFL Mobile, etc.

Video Delivery


Customer patience was tested recently when Showtime’s OTT Streaming video service saw its first On- Demand outage (Posted Jul 15, 2015 by Sarah Perez(@sarahintampa)). Showtime’s recently launched OTT service hit its first snag that night when the Showtime apps on Apple and Roku devices experienced a brief outage.

So what can be done to assure OTT video services?

First, we need to understand what type of video we are talking about.

There are still multiple video media transports in use by OTT video content providers including RTMP (Real-Time Messaging Protocol), progressive download and RTSP (Real-Time Streaming Protocol). However, this discussion will concentrate on adaptive HTTP streaming as the industry consensus is making it the de-facto solution for OTT audio-visual streaming services.

For Adaptive HTTP streaming, OTT uses HTTP-based servers from which videos are delivered as small files over TCP (Transmission Control Protocol). The leading examples are:

  • Apple HTTP Live Streaming (HLS)
  • Adobe HTTP Dynamic Streaming (HDS)
  • Microsoft Smooth Streaming

Both Adobe and Microsoft have declared that they will deprecate their proprietary schemes in favor of DASH (Dynamic Adaptive Streaming over HTTP), which is largely a combination of the two. Adobe has also declared that they will deprecate RTMP in favor of HDS and now DASH.

Adaptive HTTP streaming not only controls the rate at which content is downloaded, but also allows the client to seamlessly adapt to the transmission bandwidth available. All of the schemes listed above work in essentially the same way, as described below.

The audiovisual content is encoded in segments that are a few seconds in duration, typically 0.5 to 10 seconds, depending on the scheme. Each segment is stored on the server in a separate media file. File segments are only downloaded just before they are required, thus avoiding the progressive download problem whereby large amounts of unused content may be downloaded.

Adaptive HTTP streaming can also adapt to the bandwidth available. This is enabled by adaptive schemes that make the content available at a number of different combinations of bit-rate and image resolution. This allows the client to deliver the best quality the link will support and adapt to any subsequent changes in link bandwidth.

Some factors that inhibit the popularity of OTT include the following: Streaming videos take up large amounts of cell phone data, and since stream quality depends on the quality of the IP connection, some people still have computers too slow to stream videos effectively. In addition, while some are able to stream videos from their TVs, others must watch videos on a PC, which is less compatible with social viewing—a possible deterrent for groups such as families.

The Monitoring adaptive HTTP video streaming
The three most important parameters to extract from an adaptive video HTTP stream are the asset ID (what people are watching), the bit-rate (the higher the bit-rate), the better the video quality) and the resolution (degree of image clarity e.g. 1080p, 720p, VGA, CIF,).  Information about the client device is also useful and can be obtained from the User Agent HTTP field.

A degree of asset tracking can be provided by defining HTTP applications by URL (an example of an asset is a particular movie or channel). It is also possible to detect missing HTTP segments just like any other missing HTTP content.

The estimation of the frequency and duration of freeze events estimation is also seen as valuable for monitoring OTT video services allowing Operations and Engineering to take proactive approaches to increases in these events.

Network operators and ISPs (Internet Service Providers) are taking an increasing interest in OTT video traffic because it represents such a large portion of the data they carry. They have an interest in both the usage patterns, for capacity planning, and how well content is delivered – simply because it represents one of the main use cases of their service. Another important consideration is that some cable operators are bundling OTT services as part of their service, for example, Comcast relies on Hulu to deliver a portion of their video on demand portfolio.

OTT differs from other streaming video in many ways. Some of these differences present new challenges for content providers to overcome, but when successfully addressed, providers can create an excellent user experience.