AT&T LTE Test-Drive Results Revealed

As AT&T Inc. (NYSE: T) officially launched Long Term Evolution (LTE) in five cities on Sunday, the first performance results of the newest U.S. LTE network have been revealed and the downloads are fast.

Signals Research Group conducted a test-drive of AT&T's LTE network in Houston. For the test, the firm transferred 88.8 Gbytes of data over a three-day period, using the Sierra Wireless Inc. (Nasdaq: SWIR; Toronto: SW) USBConnect Momentum 4G dongle, and covered 205.6 miles of the Houston area.

The key findings are as follows:

  • The average downlink speed was 23.6 Mbit/s and the peak data rate was 61.1 Mbit/s, which "meaningfully exceeded" the expectations of Signals Research. The data rate was higher than 5 Mbit/s for 95 percent of the time.

  • The average uplink speed was 15.2 Mbit/s with a peak rate of 23.6 Mbit/s, which was "much higher than we anticipated," according to Signals Research. The uplink data rate exceeded 5 Mbit/s 98.2 percent of the time. [Ed note: As a shared medium, wireless networks typically slow down as they get loaded up with customers, so it will be worth checking these speeds again in a year.]

  • The average latency was 49 milliseconds with a minimum value of 40 milliseconds. According the Signals Research, this was "higher [worse] than we were expecting based on the capabilities of LTE and our experiences with testing networks in Europe."

  • For the handover between LTE and HSPA+, it took 2.4 seconds for the device to drop the LTE connection and establish a connection on the HSPA+ network, which Signals described as "relatively seamless for a data connection."

AT&T has launched LTE networks in Atlanta, Chicago, Dallas, Houston and San Antonio. The operator plans to offer LTE in at least 15 markets, covering 70 million people, by the end of the year.

Why this matters
These are the first insights into the performance of AT&T's new LTE network in the U.S. and provide an early benchmark for user expectations. As for the worse-than-expected latency results, Signals explained that the findings were on a par with other LTE networks in North America.

For more
For more on Ma Bell's LTE, please see these stories.

— Michelle Donegan, European Editor, Light Reading Mobile

DavidEHughes 12/5/2012 | 4:53:22 PM
re: AT&T LTE Test-Drive Results Revealed

How much spectrum was used or what was the average bits/hertz?  Was the system loaded with other subscribers or was the test the only user?

Alton 12/5/2012 | 4:53:20 PM
re: AT&T LTE Test-Drive Results Revealed

What user experience can be expected? At what average distance from an eNode B would the 4G connection drop out and a 3G connection take over? An indication of the extent of the LTE coverage is needed. Otherwise this DL/UL data rate information is meaningless.  In the absence of a coverage map the average DL/UL data rates for an aggregate 4G/3G connection, say for oustonan area of Houstonaan area of Houston where the LTE rollout is complete, would be helpful. 

Michelle Donegan 12/5/2012 | 4:53:18 PM
re: AT&T LTE Test-Drive Results Revealed

AT&T uses a 2x10MHz channel now.

Also, I checked with Signals Research and they were the only user on the network.




Mike Thelander 12/5/2012 | 4:53:15 PM
re: AT&T LTE Test-Drive Results Revealed

The results are based on covering roughly 210 miles of greater Houston in a 15 mile by 15 mile grid, plus a >20 mile stretch out to the airport.  The coverage is actually much greater than where the testing occurred since there wasn't time to drive the full area.

Since LTE is deployed at 700MHz on the operator's 850/1900MHz footprint there shouldn't be a handoff from LTE to HSPA+ as long as an adjacent LTE cell site exists.  In this case the handovers occurred because the network is still being built out and not all cell sites have been commissioned.

You can download the free report from the SRG website and see a geo-plot of the throughput over a Google map image.

Radhakant 12/5/2012 | 4:53:11 PM
re: AT&T LTE Test-Drive Results Revealed 40-49msec air interface delay is remarkable. Any idea until what point it was measured?
OpEd 12/5/2012 | 4:51:32 PM
re: AT&T LTE Test-Drive Results Revealed

"...relatively seamless for a data connection." 

Can users expect to have to restart their data application sessions? Was the connection maintained through the handover or was a restart required?

I wonder if this number increases when the number of users and handover events goes up...



Sign In