Network Sunsets and the IoT: What’s Happening and What You Must Do

Any historian will confirm that process does not mean progress. Simply because a technology is older does not mean that it will be decommissioned first. In a growing number of markets and service providers, it has been realized that 3G will outlive its usefulness before 2G does. We are currently seeing 3G networks being earmarked for closure before 2G across Europe—by Vodafone in the Netherlands, Telenor in Norway, and the Tele2/Telia joint venture Sunab in Sweden, as well as a possible closure by T-Mobile in the United States. The 3G-first phenomenon is not limited to “developed” markets nor to those companies (such as India’s Bharti Airtel) that were the first to deploy 3G.

The most often-stated reason for delaying the shutdown of 2G, or for the closing of 3G before 2G, is the continuing need to support and grow Machine-to-Machine (M2M) services. This fact has been acknowledged for years. While the commitment was once thought to be an Achilles’ heel for carriers, the transition of long-term M2M contracts to new network technologies can be carefully managed over time. A far greater inhibitor of 2G shutdowns (for those carriers that wish to do so and do not have a substantial 2G M2M base to maintain) comes in the form of a small number of dedicated 2G consumers.

The most commonly cited reasons why subscribers have long held on to 2G are the need to manage personal finances with a cheapest possible tariff plan; the desire not to be distracted by Internet services during the day and to reserve Internet use for leisure time at home, as well as not being interested in smartphones; and an earnest love of a companion device of many years that is almost indestructible and features an unrivaled battery life, even for today. Another notable group of die-hard 2G users comes from government bodies, who bought pools of phones for ad hoc distribution among their field force employees—most notably health workers.

In some countries, however, legacy network closure has not been due to the carriers’ need to refarm spectrum assets—it has been forced. The most notable example is Taiwan, where all 2G and 3G networks will be closed by the end of 2018 due to the expiration and nonrenewal of spectrum licenses. In these types of instances, large numbers of consumers need to be forcibly migrated, causing difficulty, disruption, and even fines for some operators for poor process management and the resulting customer dissatisfaction. This happened to Asia Pacific Telecom during the closure of its 3G Code-Division Multiple Access– (CDMA-) based network.

The decommissioning of 2G has occurred with ceremony—most notably in Australia. Earnestly projecting the trusted and dependable relationship end users and the telecoms industry cultivated with Global System for Mobile Communication (GSM). GSM is the technology that created the world of wireless communications that we live in now. GSM did not represent the advent of mobile telephony, but it was the digital service innovation that made mobility scalable and affordable enough to individually serve the population of nearly the entire world. It was the tipping point for runaway, mass-market success and is the closest we have come to a literal “worldwide” service.

In preparation for a network closure, a carrier may decide to stop accepting new device activations or to stop issuing new device certifications. This can occur 12 to 18 months before a network shuts down and even before a specific sunset date is formally announced. Stopping device activations is a tactic used to target consumer and corporate-liable markets for mobile phones. It prevents new subscribers with new stocks of old devices from joining the network. Meanwhile, ceasing to issue device certifications targets the M2M market. It dissuades module manufactures from developing new models that use legacy technologies and encourages developmental efforts to shift to newer models—thus reducing legacy module shipments and (in time) reducing the number of new devices built using legacy module stocks that appear onto the network to zero.

Even in markets where time lines to eventual network sunsets are years away, spectrum-use and base station infrastructure is being pared to the minimum required to guarantee business-critical operations. In the case of T-Mobile USA, 2G network shutdown will not occur before the end of 2019. Yet the carrier has already decommissioned most of its spectrum, retaining a small amount for M2M only, and that is operating in guard bands at (purportedly) a zero-spectrum overhead and opportunity cost. Even without a formal sunset date being established, low-traffic base stations serving legacy networks can be decommissioned when they are identified as a simple matter of course—as can be seen with China Unicom.

Formal enterprise M2M customers of carriers can be given migration plans. The most concern about 2G and legacy network shutdowns is coming from small business communities that are reliant on trackers in industries such as farming. They are private enterprises that have invested in or leased expensive connected machinery. They may be indirect customers of a carrier, but they too need clear information about 2G and legacy network closures. An early, sudden, or unexpected network shutdown can certainly negatively impact a carrier’s M2M business, as shown following AT&T’s closure of its 2G network in 2016, when T-Mobile USA openly advertised its continued support of GSM in a campaign purposely targeting disaffected M2M customers.

Carriers should be pragmatic and make legacy technology work for them, treating it as a sunk cost service with a zero opportunity cost, until the legacy technology is shut down. Decommission low-traffic RAN equipment as it is identified, and refarm as much spectrum as possible. Cease activations and certifications with plenty of time to organically run down the installed base without forcing a stop and disenfranchising customers. This way, when the “big shutdown” occurs, it is a headline event in name only, with most of the network already effectively offline. Also, be sure to plan sufficiently enough ahead to not leave any customers stranded.

Whether by choice or not, shutdowns should be at least a two-year process. A carrier who is forced by spectrum license expiration to close a network should have been prepared due to the expiration date in the contract. A smooth sunset scenario is therefore one for which the carriers should always have planned. There is no excuse for not doing so, even if only because licenses may eventually not be extended. Hurried network shutdowns—intentional in the case of AT&T or mandated during APTs)—can be truly destabilizing, forcing rapid and poorly considered investment through necessity and causing end-user consternation and confusion that will inevitably boomerang to service providers.

Do not be ashamed of supporting old technologies. Networks exist to provision services. For as long as the service opportunity is viable, so too should be the technology enabling the network. Newer is not always better. In 2015, KPN’s Chief Technology Officer (CTO) Erik Hoving famously referred to 3G as “very mediocre,” saying, “3G was from a technology perspective a half-baked data solution.” LTE and 2G complement each other well, with the best use of 3G as a source of new spectrum for LTE. The 2G technology is also vital in kick-starting the Low-Power, Wide-Area (LPWA) and IoT market for inclusion in multimode Cat-M/Narrowband IoT (NB-IoT)/GSM modules. GSM provides a fallback where cellular LPWA is currently absent to guarantee the cheapest and best coverage connectivity, wherever a device may be.

Eventual shutdown is inevitable, so connected device vendors must be alert. Even carriers that are not due to close their 2G or 3G networks for some time are likely to be paring those networks as much as possible, as soon as possible. For M2M customers especially, staying on a network until the very end will not be ideal; an earlier move than is suggested by the sunset date may be needed. Vendors of connected devices must think about the products they are launching now and in the near future, along with what the life expectancy will be. They must consider what technologies will and will not be available at that horizon’s end to know what technologies to build in right now. And if a vendor’s connectivity partner is a Mobile Virtual Network Operator (MVNO), the vendor must seek guarantees that all the right carrier agreements are in place for the technologies they will need and for when and where they will need them.