Several network initiatives that AT&T is unveiling this week show the carrier is far from finished advancing its network even as it achieves a broad footprint with LTE.
The company plans to roll out LTE Broadcast technology over the next three years to help it deliver specialized content in specific locations, among other things, Chairman and CEO Randall Stephenson told a Goldman Sachs financial conference on Tuesday morning. In the same session, he laid out AT&T's goal to convert its whole infrastructure into a system based on LTE, IP (Internet Protocol) and cloud computing, replacing traditional copper phone lines by 2020.
On Monday, AT&T signaled a commitment to SDN (software-defined networking) and NFV (network functions virtualization), including seeking out some technology vendors with new ideas. And in a nod to the potential of unlicensed Wi-Fi networks, on Tuesday it announced a partnership with Fon to expand international Wi-Fi roaming.
LTE officially stands for Long-Term Evolution, and the global standard is itself evolving into upcoming versions that will allow for features such as combining separate frequency bands into one. But AT&T, like other carriers, is also continuing to find ways to make its network more efficient and better able to handle mobile data demand that Stephenson said is rising by about 50% every year.
One way AT&T hopes to meet that demand is with LTE Broadcast, which is designed to send specific kinds of content at certain times and places, such as sports venues. This can ease the burden on a carrier's general wide-area mobile network, reducing congestion and boosting subscribers' speeds.
AT&T's plan for LTE Broadcast is a kind of homecoming for the technology, which has its roots in Qualcomm's defunct FLO TV. After the FLO TV video service that Qualcomm delivered failed to catch on, AT&T acquired the spectrum over which it ran. The carrier now plans to use those frequencies for LTE Broadcast.
The spectrum buy gave AT&T 12MHz of frequencies in the most populous areas of the U.S. near the coasts and 6MHz toward the middle of the country, Stephenson said. It will use that spectrum to deliver video and other content in settings where most people are interested in seeing the same things, he said.
As an example of how LTE Broadcast might be used, a carrier could activate it in a football stadium during a game to deliver football-related content, such as instant replays and highlight videos from other games. Instead of each fan pulling down an individual stream for each of those videos, one stream could go out to all of the subscribers in the stadium.
LTE Broadcast could also be used for live news reports or sports events over a broader area. Another use that some envision for the technology is delivery of big chunks of content, such as OS updates, during the night or times of day when the cellular network isn't as heavily used.
A few key differences between LTE Broadcast and FLO TV should make the newer technology more successful, Ovum analyst Daryl Schoolar said. For one thing, LTE Broadcast won't need a dedicated network like the one used for FLO TV, he said. The spectrum could be allocated to broadcasting at times and in places where it's likely to be needed rather than being locked into sending the same channels all the time to a whole city. When it's not needed, those frequencies might be aggregated into the block used for general mobile data, Schoolar said.
In addition, more devices are likely to be able to carry LTE Broadcast. Network vendors including Alcatel-Lucent and Ericsson are backing the technology, and Qualcomm executives said in April that the company was building support into its popular Snapdragon mobile processors.
Stephenson expects AT&T's LTE Broadcast infrastructure to be "mature in scale" within three years. AT&T's rival Verizon Wireless is pursuing the same technology, with executives there saying they plan to offer services next year.
AT&T has also committed itself to new technologies for its underlying network architecture. On Monday, the company unveiled the second generation of its Supplier Domain Program, through which the company plans its infrastructure and selects vendors to supply its parts.
In the new phase, called Domain 2.0, AT&T plans to use SDN and NFV to cut costs and offer new services more quickly.
SDN separates the control plane that governs a network from the forwarding plane that sends packets through it. NFV removes the processes underlying network services from specialized hardware devices and turns them into pure software that can run on less expensive generic computing platforms.
"The hardware is a commodity-priced product and the software is where all the intelligence is," Stephenson said. AT&T is already virtualizing its own data centers and wants to extend the process to its core network, he said.
The new approaches should help the carrier create new services and applications, generate revenue from them more quickly and deliver high performance, security and reliability on the network, AT&T said.
Both are fairly new, with standards and APIs (application programming interfaces) still emerging. But in a press release that gave few details, AT&T laid down a firm commitment to implement them and said it would start selecting vendors and signing deals for Domain 2.0 later this year and through next year.
"Over the next five years, you will see a virtualization of the network," Stephenson said.