AT&T: Delay in 3G sunset would ruin 5G rollout

ATT
AT&T cites statements from alarm industry executives who told investors that Covid has not kept technicians out of people's homes. (Getty Images)

AT&T on Monday filed its response to the alarm industry’s attempts to keep the 3G network up and running, essentially chalking it up to a delay tactic designed to line the pockets of alarm companies.

While all three of the major U.S. wireless operators are shutting down their 3G networks – and T-Mobile has garnered the lion’s share of media attention – the alarm industry is especially reliant on AT&T. That’s in part because AT&T offered aggressive pricing deals back when 3G networks were just getting started.

Fast forward to today, and those alarm companies are in no hurry to switch to newer networks. In fact, they blame Covid-19 for their inability to get into customers’ homes to do necessary upgrades. The Alarm Industry Communications Committee (AICC) filed a petition (PDF) earlier this year asking the FCC to force an extension of AT&T’s 3G shutoff, which is currently scheduled for February 2022.  

AT&T responds: No way

AT&T argues that the FCC has no legal authority to order a delay in its 3G sunset. However, even if it did, there’s no “sensible policy basis” for it to do so.

“The alarm industry provides no basis for its assertion that delaying AT&T’s 3G sunset date is necessary to prevent harm to anyone or anything apart from the alarm companies’ own profit margins,” AT&T stated.

AT&T said it is dedicated to serving customers during the network transition and has followed the same playbook it used when retiring the 2G network back in 2017. In its eyes, AT&T has given its customers ample time to do the necessary preparations, and Covid-19 is no excuse.

RELATED: AT&T 3G sunset sets off alarms for monitor industry

In fact, AT&T quotes alarm industry executives, including those from ADT and Alarm.com Holdings, talking about 2020 as a “banner year.” Despite early safety protocols that prevented technicians from entering homes, signs of recovery started as early as May 2020, when technicians were installing residential alarm systems at a “record pace, exceeding even pre-pandemic forecasts,” according to the AT&T filing.

“Indeed, several alarm companies credited the pandemic for accelerating new alarm installations,” AT&T told the commission. 

The alarm industry has suggested that AT&T delay its 3G network shutdown to the same timeline Verizon has given: December 2022, but ATT said if it were forced to delay the 3G sunset, it would impair AT&T’s 5G rollout and degrade network performance “to the detriment of tens of millions of American consumers.”

Reshuffling the 850 MHz deck

In the filing, AT&T spelled out what’s going on with its 850 MHz spectrum as it relates to the refarming of it for 5G. Kevin Hetrick, VP of Access Construction and Engineering, told the commission that AT&T’s nationwide 5G coverage currently relies on low-band spectrum – specifically, the 850 MHz blocks that aren’t being used for 3G.

In the future, the operator will rely mostly on mid-band spectrum and in particular the C-band assets that it acquired in Auction 107, which raised a record $81.2 billion. However, the C-band spectrum isn’t yet available, and in the meantime, AT&T will be converting the 850 MHz spectrum, currently used for 3G, to 5G for both coverage and capacity.

Here’s how it currently breaks out. AT&T devotes a narrow 10-megahertz sliver of 850 MHz spectrum to its 5G network, in the form of paired 5x5 channels – five megahertz for downlink and five megahertz for uplink. While that’s sufficient for today’s 5G traffic, it’s not going to be enough to support the expected demands in 5G for 2022, according to Hetrick.

Two crucial steps are involved. First, by year-end 2021, AT&T will begin to expand it 5G network capacity by adding C-band (3.7 GHz) spectrum in increasing numbers of markets. That’s not going to be enough, however. Because C-band propagates over shorter distances than 850 MHz spectrum, customers at the edge of the C-band coverage areas will still need to rely on the 850 MHz band for uplink 5G transmissions. But AT&T’s existing 5 MHz uplink channel in the 850 MHz band won’t be enough to meet the expected traffic demand in 2022.

That’s where the second step kicks in. AT&T plans to double its allocation of 850 MHz spectrum to 5G by repurposing the 10 megahertz of spectrum now being used for 3G. Combining the 10 megahertz of refarmed 850 MHz spectrum with the 10 megahertz of 850 MHz already used for 5G will mean 20 megahertz dedicated to 5G, in the form of 10x10 megahertz (10 uplink and 10 downlink) channels.

Of course, none of this can happen if AT&T is forced to keep using this spectrum for a 3G network, which it deems “obsolete.” Failing to upgrade AT&T’s current 5x5 megahertz channels of 850 MHz spectrum into 10x10 megahertz channels would increase the rate of busy-hour “blocking” in cell sectors across virtually all geographic markets and result in more dropped and blocked calls with lower data throughput. “Such network degradation could affect tens of millions of customers over the course of 2022 if AICC’s petition is granted,” AT&T told the commission.

RELATED: Verizon sets final deadline for 3G sunset

In explaining why AT&T can’t delay its 3G shutdown in the same way that Verizon has done, Hetrick explained that the two companies use materially different 3G technologies. Verizon uses CDMA/EVDO for 3G, which can operate with paired 1.25 megahertz channels – far narrower slices of spectrum than the 5x5 paired to serve even a single UMTS customer on AT&T’s network. For that reason, “in many markets Verizon would be able to dedicate a full 20 megahertz of 850 MHz spectrum to its LTE or 5G network while still supporting 3G legacy users on its CDMA/EVDO network. Again, AT&T does not have that option because of the minimum spectrum needs of UMTS.”

In other words, unlike AT&T, “Verizon appears capable of dedicating the lion’s share of its 850 MHz spectrum to 5G while still supporting legacy 3G users,” according to the filing.