Policy driven optimization

The video optimization market is still young, but with over 80 mobile networks deployed globally, I am officially transitioning it from emerging to growth phase in the technology life cycle matrix.


Mobile world congress brought many news in that segment, from new entrants, to networks announcements, technology launches and new partnerships. I think one of the most interesting trend is in the policy and charging management for video.


Operators understand that charging models based on pure data consumption are doomed to be hard to understand for users and to be potentially either extremely inefficient or expensive. In a world where a new iPad can consume a subscriber's data plan in a matter of hours, while the same subscriber could be watching 4 to 8 times the same amount of video on a different device, the one-size-fits-all data plan is a dangerous proposition.


While the tool set to address the issue is essentially in place, with intelligent GGSNs, EPCs, DPIs, PCRFs and video delivery and optimization engine, this collection of devices were mostly managing their portion of traffic in a very disorganized fashion. Access control at the radio and transport layer segregated from protocol and application, accounting separated from authorization and charging...
Policy control is the technology designed to unify them and since this market's inception, has been doing a good job of coordinating access control, accounting, charging, rating and permissions management for voice and data.


What about video?
The diameter Gx interface is extensible, as a semantics to convey traffic observations and decisions between one or several policy decision points and policy enforcement points. The standards allows for complex iterative challenges between end points to ascertain a session's user, its permissions and balance as he uses cellular services. 
Video was not a dominant part of the traffic when the policy frameworks were put in place, and not surprisingly, the first generation PCRFs and video optimization deployments were completely independent. Rules had to be provisioned and maintained in separate systems, because the PCRF was not video aware and the video optimization platforms were not policy aware.
This led to many issues, ranging from poor experience (DPI instructed to throttle traffic below the encoding rate of a video), bill shock (ill-informed users blow past their data allowance) to revenue leakage (poorly designed charging models not able to segregate the different HTTP traffic).


The next generation networks see a much tighter integration between policy decision and policy enforcement for the delivery of video in mobile networks. Many vendors in both segments collaborate and have moved past the pure interoperability testing to deployments in commercial networks. Unfortunately, we have not seen many proof points of these integration yet. Mostly, it is due to the fact that this is an emerging area. Operators are still trying to find the right recipe for video charging. Standards do not offer guidance for specific video-related policies. Vendors have to rely on two-ways (proprietary?) implementations.


Lately, we have seen the leaders in policy management  and video optimization collaborate much closer to offer solutions in this space. In some cases, as the result of being deployed in the same networks and being "forced" to integrate gracefully, in many cases, because the market enters a new stage of maturation. As you well know, I have been advocating a closer collaboration between DPI, policy management and video optimization for a while (here, here and here for instance). I think these are signs of market maturation that will accelerate concentration in that space. There are more and more rumors of  video optimization vendors getting closer to mature policy vendors. It is a logical conclusion for operators to get a better integrated traffic management and charging management ecosystem centered around video going forward. I am looking forward to discussing these topics and more at Policy Control 2012 in Amsterdam, April 24-25.

Mobile video QOE part II: Objective measurement

Objective measurement of video is performed using mathematical models and algorithm measuring the introduction of noise and the structural similarity of video objects.

There are several mathematical models such as PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity) for instance, that are traditionally used for these calculations. The complexity resides in the fact that a mathematical difference from one pixel to another, from one frame to another does not necessarily translate equally in the human eye.

PSNR is a measure that has a medium to low accuracy but is quite economic in computation. It represent possibly up to 10% of the CPU effort necessary to perform a transcoding operation. This means that although it provides a result that is not fully accurate, the model can be used to compute calculations as the file is being optimized. A vendor can use PSNR as a basis to provide a Mean Opinion Score (MOS) on the quality of a video file.

Video quality of experience measurement can be performed with full reference (FR), reduced reference (RR) or no reference (NR).

Full Reference

Full reference video measurement means that every pixel of a distorted video is compared to the original video. It implies that both original and optimized video have the same number of frames, are encoded in the same format, with the same aspect ratio, etc… It is utterly impractical in most cases and requires enormous CPU capacity to process, in many cases more than what is necessary for the actual transcoding / optimization.

Here is an example of a full reference video quality measurement method under evaluation and being submitted to ITU-T.

As a full reference approach, the model compares the input or high-quality reference video and the associated degraded video sequence under test. 

Score estimation is based on the following steps:

1)          First, the video sequences are preprocessed. In particular, noise is removed by filtering the frames and the frames are subsampled.

2)          A temporal frame alignment between reference and processed video sequence is performed.

3)          A spatial frame alignment between processed video frame and the corresponding reference video frame is performed.

4)          Local spatial quality features are computed: a local similarity and a local difference measure, inspired by visual perception.

5)          An analysis of the distribution of the local similarity and difference feature is performed.

6)          A global spatial degradation is measured using a Blockiness feature.

7)          A global temporal degradation is measured using a Jerkiness feature. The jerkiness measure is computed by evaluating local and global motion intensity and frame display time.

8)         The quality score is estimated based on a non-linear aggregation of the above features.

9)          To avoid misprediction in case of relatively large spatial misalignment between reference and processed video sequence, the above steps are computed for three different horizontal and vertical spatial alignments of the video sequence, and the maximum predicted score among all spatial positions is the final estimated quality score.

Reduced reference 

Reduced reference video measurement is performing the same evaluation as in the full reference model but only on a subset of the media. It is not widely used as frames need to be synchronized and recognized before evaluation.

No reference 

No reference video measurement is the most popular method in video optimization and is used usually when the encoding method is known. The method relies on the tracking of artefacts in the video, such as blockiness, jerkiness, blurring, ringing…etc. to derive a score.

Most vendors will create a MOS score from proprietary no reference video measurement derived from mathematical models. The good vendors constantly update the mathematical model with comparative subjective measurement to ensure that the objective MOS score sticks as much as possible to the subjective testing. You can find out who is performing which type of measurement and their method in my report, here.