The case for switching off 2G and/or 3G networks is becoming more compelling, with operators able to more efficiently deploy 4G and/or 5G in the spectrum bands they once occupied. Phasing out legacy technologies may also be needed on the migration to OpenRAN and the move away from single RAN solutions.
Approaches to the switch-off of 2G and/or 3G vary between countries. In some (Australia, US, Canada) they are almost entirely unregulated, whereas in others (Japan, South Korea) regulators oversee the process to protect consumers. In either case, the process is generally initiated by the operators rather than being driven by the regulator or policymakers.
Across Australia, Canada, Japan, South Korea, and the US, 2G is now almost entirely a thing of the past. 13 out of 16 operators in these countries phased it out between 2008 and 2020. While 3G networks are still active in these countries, seven out of 16 operators have announced plans to sunset it, with the US set to become the first country where 3G will be retired entirely, once Verizon phases it out in 2023.
Communication campaigns played an important role in regulators’ decisions to authorise network sunsets. Communicating extensively (some operators even visited the most vulnerable customers in person) and well in advance ensures that regulators feel less of a need to intervene in the process. On average, 2G sunsets have been announced about three years ahead of implementation. Operators have been more cautious with 3G, for which the average notice has been about four years so far.
In some cases, regulators made sure consumers were given adequate incentives to migrate to newer technologies, even preventing switch-off until they had. On the other hand, there has been less attention paid to M2M users, which have had to deal with network sunsets without the protection of regulators.
Why it’s time to start thinking about 2G/3G retirement
Despite the near-ubiquitous availability of 4G (91% of households in the UK in 2020, and 96% in the EU) and the accelerating pace of 5G deployment, we can’t yet consider 2G and 3G a thing of the past. In the UK, 4G made up 75% of all connections in 2019, meaning about a quarter were still 2G or 3G. The GSMA reports that 2G and 3G connections made up 42% of all connections in Europe in 2019, and expects that they will still be about 8% in 2025. But the time for the retirement of those networks is coming, and in some countries operators have already phased out at least one of their legacy networks, or have made plans to do so.
One driver for the acceleration of phasing out legacy technologies is the opportunity to deploy 4G and/or 5G in the spectrum bands that are occupied by 2G and/or 3G. This allows operators to simplify network management, avoids costly maintenance of ageing infrastructure, and reduces the need for (and the cost of) tower space for multiple antennas. Not to mention the reduced energy consumption of a network, which is even more important at a time when operators are getting serious about their approach to sustainability.
Simplified networks are also a more attractive proposition for network vendors – especially in view of the ongoing diversification of the supply chain that many countries are carrying out in their transition to 5G. In the UK, the Government’s Telecoms Diversification Task Force makes clear that single RAN solutions like those currently adopted by UK operators require vendors to support old technologies as well as the new ones. This would make it harder for the UK to attract new vendors, which may not be able or willing to support technologies that are being phased out around the world. This is particularly important as operators embrace OpenRAN, whose ecosystem includes small vendors.
The phasing out of 2G and/or 3G has largely been operator-led with varying degrees of regulatory intervention
Approaches to sunsetting 2G and/or 3G networks have varied between countries, from purely market-led initiatives (US, Canada, Australia) to close regulatory oversight (Japan, South Korea). Despite the different approaches, the initiative to phase out 2G and/or 3G networks has generally been left to the operators. Where regulators have intervened, it has been to avoid or minimise disruption for consumers. If the UK Government takes on board the advice of the Task Force, the UK will have a sunset process that is mainly government-driven, since operators have so far been silent on their plans to phase out these older networks.
In the US, the sunsetting of 2G has had no push from the regulator, the FCC, and the same will be true with 3G. The FCC is clearly aware that sunsetting 3G will affect a significant number of Americans (currently more than 43m users) and that it goes beyond mobile phones (since security systems and medical devices are affected too). Despite the scale of the problem, the FCC has only issued minimal guidance to consumers, advising them to plan ahead and to talk to their provider. Network sunsets have even triggered competitive dynamics between operators. In 2016, T-Mobile marketed its 2G network as a ‘lifeline’ for AT&T customers still reliant on the technology, and promised to keep it live until 2020.
Similarly, in Australia, decisions have been left to the operators. Telstra, Optus, and Vodafone phased out their networks in December 2016, August 2017, and June 2018 respectively. The communications regulator, the ACMA, has taken a hands-off approach and let industry deal with consumers directly. However, once the 2G switch-off was complete, the ACMA took the opportunity to undertake a comprehensive review of the use of the 900MHz band, which has been reconfigured to support deployment of both 4G and 5G. The Government did intervene in one instance to ask Telstra to delay the switch-off of the CDMA network, which Telstra had wanted to phase out between 2007 and 2008. Only in April 2008 did the Government authorise them to do so, once the ACMA had verified that Telstra’s new 3G network had reached an equivalent level of coverage. Of all the operators, only Telstra has announced detailed plans to phase out 3G, which has triggered dynamics similar to those we have seen in the US.
Conversely, in Japan and South Korea, regulators have sought greater control of the process. This is mainly for two reasons: the fact that spectrum licences were not technology-neutral (meaning that different uses needed to be re-authorised), and to minimise disruption for consumers. In Japan, the process to phase out 2G was coordinated by the Ministry of Internal Affairs and Communications (MIC) but came after requests from operators. Regulatory coordination was also necessary in light of the refarming of the 800MHz band, which was initially announced by the MIC in 2004. At the time, NTT DOCOMO and KDDI were using this band for both 2G and 3G. The switch-off of 2G facilitated the process of reducing fragmentation and reassigning the spectrum to MNOs.
The switch-off processes for 2G and 3G in South Korea have been closely monitored by regulators (initially by the KCC, now the Ministry of Science, ICT and Future Planning (MSIP)). In theory, once permission is granted operators can stop services after a relatively short notice (just 60 days). However, the MSIP only grants approval once the user-base for a given technology has become very small (around 1% of an operator’s total subscribers), as was the case for KT’s sunset of 2G in 2012, for which the operator had to submit multiple requests for authorisation during 2011. SKT faced a similar process between November 2019, when it first submitted its request to phase out 2G, and June 2020, when the MSIP finally approved the request. Operators must show they have measures in place to protect end-users, including MVNO subscribers hosted on their networks.
In all five countries we have studied, the sunset of 2G has occured before the phase out of 3G. In these countries 13 out of 16 operators have already phased out 2G. The phasing out of 3G is coming too, although only seven out of 16 operators have announced it so far, and the earliest scheduled date is not until next year. Table 2 shows how things compare operator-by-operator.
Operators engaged in extensive communication campaigns, not only where regulators required them to do so
There is something of a mixed picture when it comes to how well operators have prepared their customers for a network retirement. Sometimes they have given ample notice even without regulatory pressure, others they have tried to do the bare minimum which has tended to trigger regulatory intervention. In particular, regulatory scrutiny has led operators in Japan and South Korea to communicate their 2G phase-outs extensively.
In the US, where things were more left to the industry, operators too have given ample notice of their plans to phase out 2G. For example, AT&T announced it in 2012 – more than four years ahead of executing it in 2017. On top of the lengthy notice, the company said it engaged in frequent communication with customers to encourage their migration to newer technologies (it is understood that AT&T also sent letters to its customers several months ahead of the phase-out). Similarly, Verizon told its customers about the 2G sunset in 2016 – more than three years ahead of its eventual phase-out at the end of 2019. There has been less clarity about T-Mobile’s 2G sunset plans, initially foreseen for early 2021 and now reportedly delayed to 2022. However, it is likely that customers and businesses by now have developed a higher level of readiness and acceptance of the fact that 2G is being retired, and should be more prepared than they were four or five years ago. Similarly, AT&T and Verizon announced their 3G sunsets well in advance (about three years ahead of the phase-out, which will take place in February 2022 and January 2023, respectively). However, AT&T’s communication of the process has been more confusing than for 2G. In 2020, AT&T’s announcements led customers to believe that the operator was phasing out its 3G network immediately. The operator rectified its message and apologised.
Canadian operators have generally given fair notice of the phase-out of 2G networks to customers, and appear to have been mindful of the needs of consumers and IoT users. In particular, Rogers repeatedly postponed its GSM sunset to 2021, and sent letters to customers to keep them up to date with its plans. Initially it was announced in 2012 and scheduled for 2018, however it is now expected to take place by the end of 2021, in the end essentially giving consumers a 9-year notice period.
Australian operators are also giving plenty of warning of the phasing out of 2G and 3G networks – perhaps learning from the conflict between Telstra and the Government around the sunset of the CDMA network in 2008. At the time, Telstra’s move left a significant number of customers worried that they would be left without a network. In 2014, Telstra announced its 2G sunset almost two years ahead of implementation, noting that it had not sold a 2G device for several years. However, it is understood that the operator took care of its last few 2G customers ahead of the sunset date, communicating directly with them, sending care packages with a replacement device and even inviting the last 2G customer to personally turn off the network. For its 3G sunset (which will take place in 2024), the operator has been even more careful, with almost five years’ notice. The operator is mindful of the needs of enterprises and M2M devices, but believes that almost all IoT use cases and devices will have a 4G migration solution within 24 months.
In Japan, all three operators engaged in campaigns to proactively inform customers about the sunsets, which happened at different times. Advertising, direct mail, and information provided in shops were the main ways in which the closures were communicated. Months before announcing the switch-off, KDDI had also announced it would stop accepting new 2G subscribers, which in practice acted as an additional notice for customers to prepare for the transition. All operators are yet to switch off their 3G networks, but they have already announced their intention to do so and have given ample notice to consumers (even as much as six years in the case of DOCOMO, which will phase out 3G in 2026).
Similarly, in South Korea, communication campaigns played an important role in regulators’ decisions to authorise the 2G sunsets of KT and SKT. In particular KT widely publicised the sunset in national newspapers, made phone calls to the subscribers affected, and even visited the most vulnerable customers in person.
Despite strong communication campaigns, dealing with the migration has posed some challenges
Since sunset decisions have been mainly commercially driven, there has generally been no compensation for the operators to phase out legacy networks. Where regulators have set criteria to be met for the switch-off to go ahead, operators have often compensated consumers in some form – e.g. by offering subsidised switching to other technologies or operators, including free devices. This has been the case in Japan and in South Korea, where operators introduced incentives for customers to migrate to 3G, such as free devices, discounts, or waiving subscription fees. In the case of KT in South Korea, these efforts did not stop about 900 customers from filing a class action lawsuit, although the Seoul High Court sided with the operator and with the KCC, noting that the damages of the sunset were not irreparable for customers.
Some problems have emerged in the US, where operators have been in complete control of the sunset process. Some industries reliant on 2G networks, such as alarm communications systems, have met the costs of upgrades themselves, because they could not rely on regulatory safeguards and were reluctant to pass on these costs to the consumer. Some transport systems experienced technical difficulties – for example, the prediction of arrival times for buses in San Francisco, where the authority said it was hoping to have more time to switch its equipment to 3G. T-Mobile’s plans for 3G appear to be causing significant disruption. The company is reportedly telling its customers that the UMTS network will close on 1 October 2021, followed by the CDMA network on 1 January 2022. Rural wireless operators have asked the FCC to intervene, noting that the closure is already underway and hinders some customers’ ability to roam when they leave their home networks. The plan is also having a disruptive effect on Dish, which bought Boost Mobile as part of the T-Mobile/Sprint merger, and was planning to use T-Mobile’s 3G network to host half of Boost Mobile’s customers until July 2023.
