LTE not getting a chance to strut its stuff, report says

Mobile video may be the king of data hogs, but Signals Research Group contends that other types of data traffic--such as update pings generated by Facebook and other "chatty" applications that constantly query the network for updates--are actually consuming a disproportionate amount of network resources. Not only that, these types of data traffic do not take advantage of LTE features such as MIMO, meaning operators and vendors need to look elsewhere to enable more efficiencies, the firm said.

"It's very inefficient for LTE to deliver that data because all the bells and whistles of LTE, MIMO, using higher modulation--sending more bits in a limited amount of time--all of those things only work well when you have a lot of data to send. When you have less data to send, you can't use those capabilities. It's overkill," Michael Thelander, Signals Research's founder and CEO, told FierceWirelessTech.

In a new report, Signals Research analyzed information collected during LTE network testing at locations in San Francisco and Phoenix, Arizona, including the Arizona State University campus in Tempe. The research included all four national networks: AT&T Mobility (NYSE: T), Sprint (NYSE: S), T-Mobile US NYSE:TMUS) and Verizon Wireless (NYSE: VZ). The networks used infrastructure from Alcatel-Lucent (NYSE: ALU), Ericsson (NYSE: ALU), Nokia (NYSE:NOK) Solutions and Networks as well as Samsung.   

This chart shows the median and maximum number of simultaneous connected devices that were observed on the monitored networks.

The tests employed Sanjole's WaveJudge 4900A LTE Analyzer and IntelliJudge test platform to capture and analyze data from the physical layer and other protocol exchanges that occurred over the air between an eNodeB and active mobile devices within its coverage area. The point of the testing was to examine how users really use cellular service and what impact that has on the network, Thelander said.

There might be hundreds of devices connected to a cell sector at any given moment. "The 'problem' is that these connections are generating very little in terms of actual data traffic, and as a result the data packets are being transmitted relatively inefficiently--potentially no MIMO and low modulation rates," Thelander said.

MIMO utilization rates, or the percentage of time devices use MIMO, is surprisingly low, he noted. That can be attributed in part to the RF frontend of the antenna design in particular devices. But the use of applications that do not consume lots of data also means eNodeBs will not use MIMO by design, simply because there is not a lot of data to deal with. In addition, use of mobile devices deep within the bowels of buildings, such as high-rises, obviates the use of MIMO even for video consumption.

"If you're sitting in the bathroom stall on the 40th floor trying to watch a YouTube video, it's probably not going to happen," Thelander said.

Ultimately, because MIMO and other LTE capabilities are not being applied all the time, "LTE is delivering far less spectral efficiency than people think it does," Thelander said.

One could argue that the typical user experience is not impacted by network loading from small data transactions, given that the usual mobile data connection equates to very low data rates. However, peak data rates are being impacted by network loading, Thelander said.

He argues that inefficient use of network resources in support of modest amounts of data traffic leaves far fewer resources available to support bandwidth-intensive applications, such as video. And because small data transactions are negatively impacting overall spectral efficiency, carriers are running out of spectrum more quickly than expected.

The Signals Research report says operators must recognize that LTE alone is not a panacea and recommends that they improve network performance through additional approaches. Those include adding small cells or distributed antenna systems (DAS) for in-building coverage. Operators also should use self-optimizing network (SON) algorithms and detailed network monitoring to ensure their networks are running at their full potential.

Added the report: "Given our results, it also wouldn't hurt to have a lot more spectrum on reserve."

Related articles:
Google's Moto X tops LTE network connectivity test
Interference and the beer analogy
Macrocells struggle to deliver indoor coverage
Chatty smartphones stressing networks