Q & A: LTE doesn't obviate need for network optimization
LTE networks are more efficient than their predecessors, offering lightning-fast download speeds and improved latency. Given these improvements, you might not think that operators would need to worry about congestion on their next-generation networks.
Not so, says Bytemobile, which specializes in clearing up clogged networks with video optimization technology. The company says network optimization will be just as important for LTE networks as it is for 3G networks.
Bytemobile marketing chief Ronny Haraldsvik recently spoke about the state of mobile video traffic, network optimization techniques and what this all means to the end user. Below is an edited transcript of the discussion.
Wireless Week: Bytemobile told Wireless Week in February that the rise in mobile video traffic predicted by Cisco will happen a lot sooner than they forecast. Do you still think that's the case? What levels are you seeing in terms of mobile video traffic?
Haraldsvik: We don't only think it's the case, we know it's the case. Clearly, Cisco's Visual Networking Index is looking at worldwide traffic, but if it pinpoints a region or even pinpoints a larger operator, we're already seeing 60 percent of traffic coming from video in many of these networks. Bytemobile has insight into 36 networks.
Even a slight increase in the quality of the devices and the transmission of video has a dramatic impact. When you get to resolutions of 480, 720, 1020 pixels, you're growing traffic by a factor of four to six times.
We've done modeling and traffic analysis of existing customers, including an operator that has our solution deployed in its LTE network. 3G operators will see about a 36 percent savings in freed-up capacity by optimizing their video traffic, and as we go to LTE networks, the savings with freed-up capacity increases dramatically. We actually do better in an LTE environment.
WW: Building on what you said about LTE, could you go into a little more detail about why should an operator getting ready to deploy LTE invest in optimizing their legacy wireless networks?
Haraldsvik: Optimization, in essence, means being able to get more out of what you already have without degradations in quality. It has more to do with traffic management. Whether you're transitioning away from a legacy network, you keep that network for quite some time, and when you move to LTE or HSPA+, all that means is you're opening up your pipe for more content and for more subscribers to do more things on your network. It doesn't make the capacity problem go away. It almost makes it worse.
If you look at the advertising Sprint, Clearwire, T-Mobile and Verizon are pushing in the United States, you'll see they're conditioning subscribers to access video and rich content. That means that more, higher-quality content is going to go across the network. You have to make sure you have policies in place that shape access and bandwidth in a way that a single person can't take up all available bandwidth.
WW: LTE networks are both significantly more efficient than their predecessors and also have some built-in intelligence. Do you think there will be demand for mobile optimization services like Bytemobile's on LTE networks given their inherent efficiencies?
Haraldsvik: Absolutely. Our market increases dramatically because of LTE. Even though LTE is viewed as smarter than its predecessors, we're talking about more than putting a probe into the radio access network – all those probes do is detect. They're smart, but not smart enough.
Smart means you can do something about it, and that's where our platform comes in. You need to be able to not just detect, but react and do something about it. You need to set policies in such a way that subscribers get the right kind of content without a detrimental impact on the rest of the subscribers in that cell. I've been living LTE because I was part of Flarion, and then part of Qualcomm. Radio access networks and advanced OFDM networks – I know them cold. All it means are more lanes on a highway. You actually open up for more traffic.
WW: As you said, more lanes on a highway. Have you had to modify your optimization techniques for LTE networks versus the technology you used for 3G networks? Is the way you optimize traffic over LTE different than how you do it over 3G?
Haraldsvik: No. I guess you could release Unison 6.LTE and market it, but it would be the same platform. That's the beauty of it. We're RAN agnostic. The more traffic that gets put in our path, the more we can help the operator. Our data reduction models show we can free up anywhere between 30 percent to 50 percent of network capacity.
In one model, we had 60 percent of capacity freed up in an LTE environment. That's significant. You can get more subscribers on the same network and give existing subscribers a better, more efficient experience. Which, of course, has a dramatic impact on the reduction of churn and improving customer satisfaction.
WW: You've said that spending more to build out capacity is "a never-ending capex treadmill for operators." Could you explain what you mean by that?
Haraldsvik: Operators right now are in a very precarious position because LTE is raw bandwidth. If you just toss raw bandwidth at the problem, you'll never stop spending money on capex. The cost of capital is extremely high. This is not like going and asking for a preferred mortgage. When you build up your capex, you build up your opex. If you add 30 percent capacity to your network, that means hundreds of millions of capex in year one, adding up to billions over a two- to three-year period.
WW: It doesn't sound like you think additional capacity is the solution operators should really be going for. I'm trying to understand why you don't think "raw bandwidth," as you put it, isn't going to be enough to solve this capacity issue.
Haraldsvik: Raw bandwidth, in combination with optimization, in combination with Wi-Fi offload in some metropolitan areas, in combination with DAS and femtocells for in-building coverage – that's the combination.
If you just say across the board that you're going to spend in RAN, you're going to spend billions and get additional capacity. If you spend half that in raw bandwidth and sprinkle in femtocells, Wi-Fi offload and some optimization, not only have you reduced your capex, but increased capacity more than if you had just done raw bandwidth. We're saving operators hundreds of millions of dollars. We give operators breathing room. Before, we were nice to have. Now, we've become critical in terms of capex planning. That's a great position to be in as a company.
WW: What techniques does Bytemobile employ to optimize wireless networks – are we talking just video compression here, or does it go beyond that?
Haraldsvik: It goes way beyond that, but that's what gets the attention nowadays. We can also inspect Web traffic and do something about it. When you put Web and video together, we are indeed steering, managing, detecting and reacting to about 80 percent of the traffic running over a network. We're not just taking it back and passing it along like GGSN [Gateway GPRS Support Node]or DPI, which finds problems and passes them along. We're actually doing something about it.
WW: I'm trying to understand how this boils down from a customer point of view. If you're a wireless subscriber and your operator hires Bytemobile to optimize their traffic, what difference does that make from the end user's point of view? Will they see a difference?
Haraldsvik: Yes. They will see a big difference. It gets better.
What we're doing to content is invisible to the naked eye. Demonstrations, like what we did at Mobile World Congress, showed that roughly 50 percent guessed right, and the other 50 percent guessed wrong because it was very hard to tell what was optimized and what was not. When we showed the data to the operators afterwards, content got optimized by between 30 percent to 40 percent and reduced traffic accordingly, but the quality is the same.
It needs to be a transparent experience. A user that's not complaining is a user that's happy. The benefit of what we call quality of experience goes up as the impact on the network goes down. It's a win-win for operators, as well as subscribers.