Tekelec's recent white paper, "Policy control and mobile video: Options for managing growth" discusses the use of policy control (PCRF) to manage the delivery of video over mobile networks, with an emphasis on balancing the delivery of video to users with the impact on the network. According to Randy Fuller, director of strategic marketing at Tekelec, subscriber choice is a key component of finding that balance. In this Q&A interview, Fuller provides some insights on PCRF and the future of mobile video.
TelecomEngine: For a few years now, we've been hearing about video -- especially mobile video -- being a "bandwidth hog," and the terminology implies that the consumer is wasting valuable resources with frivolous activities (like watching cat videos). But obviously, the industry would supply the bandwidth to meet their demand if it was able, because it represents such a great potential revenue stream. Why can't the industry keep up with the demand for bandwidth?
Randy Fuller, director of strategic marketing: Video, whether frivolous or not, is a bandwidth hog because it just consumes a great deal of bandwidth relative to most other internet services. If you think about watching a video for a minute versus reading a web page for a minute, you're constantly downloading more information; because it's moving images rather than static content.
So you get situations like Netflix where, according to one of our partners, Sandvine , Netflix video streaming already represents 20% of all peak hour bandwidth traffic in the U.S. despite being a relatively young and small service. Those are real movies, not just cat videos.
On mobile networks, where delivering a bit requires spectrum, cell towers, radios, backhaul and other equipment, bandwidth availability is dependent upon operational issues plus the subscribers’ willingness to pay. Both issues are under pressure right now, and will be sorted out over the next few years.
Have you seen that people are really using mobile video in a business context today? Will usage grow, or is the demand mainly consumer-focused?
Fuller: While video for business is growing, for applications such as video conferencing and training sessions, the highest demand is still consumer.
Does policy and charging rules function (PCRF) distinguish between business use and frivolous use of mobile video? If so, how?
Fuller: No it doesn’t. The PCRF looks at things like the subscriber profile, the device type, the network state and so on, which can tell you, for example, that this is a business user or a high-value customer, versus someone who is on a different tier of service.
It does not distinguish service based on the type of video being downloaded. We don't look at the video and say, "There's a cat in this video, so don't do it," versus a video of somebody watching an interview with a CEO. It's based on whether this subscriber is a high-value customer in some way.
So you're saying that how much money the customer is willing to pay determines how much bandwidth they're allowed?
Fuller: In a lot of places, especially in Europe, absolutely. Our research  indicates that not surprisingly, subscribers value different services differently. So the segmentation that is common to other consumer goods is coming to mobile data. If it is worth someone’s money to watch a cat video, then great, let them do that.
I understand that during periods of congestion when video is consuming too much network resources, an operator can throttle video traffic on the congested nodes. How does this experience pan out for the end users? Are they understanding, or is there a backlash that the operator has to deal with?
Fuller: Typically what happens right now if you don't do anything is that everybody's video suffers.
By and large, one of the things about video is if there's some congestion, you notice it. If your web page is a little slow to load it's no big deal, but in a video it's going to freeze and chop and buffer and wait. So there's a backlash already. The question is, do you allow subscribers to make a choice how they want to be treated in those situations? In our opinion, that’s an opportunity for operators to let customers choose what is valuable to them in exchange for more or less money. Vodafone has presented data that shows when traffic controls were introduced based on subscriber tier, overall satisfaction for everyone went up since people were able to make their own trade-offs.
Is there some point at which policy control is just a stop-gap measure, like a traffic cop on a congested highway? It seems like at some point we just need more bandwidth.
Fuller: That's true. But for the financial and operational reasons I mentioned earlier, the network can only go so quickly, so the traffic cop never goes away.
If you look at the history of the Internet, if you make more bandwidth available, more people will use it. Policy control is like a traffic cop that allows operators to make the most of the bandwidth they do have – both in terms of cost savings and revenue creation. And operators will make more bandwidth available as quickly as they can afford to.
Besides mobile video, what are some other applications of policy control that are useful to operators?
Fuller: Generally, we help operators manage general Internet access: how much you can use, where you can use it, how much it costs, how much bandwidth you're allowed to use at a time, etc. The reason mobile video gets treated differently is that it's such an outlier in terms of how much bandwidth it consumes. Our research indicates it's only the fourth most important application behind Web browsing, email and navigation whereas it consumes 30% to 50% of overall bandwidth. That mismatch is why mobile video is often treated as an individual service.
Are there any notable security or privacy issues in delivering mobile video? Can policy control help, or does it actually complicate these issues?
Fuller: There are no extra security or privacy concerns using policy control because we don't know what you're watching. If there's a box examining the content of the video, then there are potential security and privacy issues. So vendors of those types of devices have to be very careful to follow the rules about that.
Your white paper  points out that "video can be treated separately from standard data traffic by the use of a separate IP traffic filter and potentially separate radio bearer." Is this common practice today or just a best practice suggestion?
Fuller: It's a best practice. Most Internet video is just delivered as part of your Internet traffic; it's just another bit. But if you want to treat it differently, as better or worse, then traffic filters and radio bearers could deliver that.
Care to make a prediction? If video traffic is taking the place of voice calling today as a carrier revenue stream, what do you see as major revenue stream potential for operators in the future?
Fuller: Video's not really a substitute for voice. Voice still makes up 70% of operators' total revenue. The fact of the matter is that voice revenue is flat, and video and data revenue is growing. So what you're seeing is that the most potential for growth in operators' bottom lines is in data traffic, in part fueled by video.
Tekelec published a paper about how consumers feel about video. The rub of mobile video is that it's a double-edged sword. It's part of what's driving people to buy iPhones and tablets, but because it consumes so much bandwidth it also puts tremendous pressure on the network. That will have to be resolved as video traffic grows.
Subscriber choice is the thing we're really trying to drive home. Mobile video is an important example of how to make that work. Subscribers ought to be willing to pay a little extra for a trade off. It's an opportunity for a win-win scenario to be able to control something that's causing them some pain, or pay the extra freight if it is worth it to them.