GSM networks in Western Europe are widely expected to be switched off around 2018–2020, but mobile operators could potentially accelerate these timescales by rapidly rolling out nationwide UMTS900 networks.
The main benefits of bringing forward the switch-off of the GSM network would be to simplify network management and realise cost savings. Maintenance of a GSM network in addition to UMTS adds a significant overhead. GSM equipment is older, takes up more space and uses more power than UMTS, and its elimination would not only enable substantial opex savings but also help with the green mandate. Moreover, maintaining the two networks in parallel would require potentially 10–15% more staff and an extra 20–30% in training opex for new starters. Further financial benefits may be derived from reselling refurbished GSM equipment to the developing world. The earlier an operator decommissions its GSM network, the greater such gains are likely to be.
Operators in markets where the GSM frequencies are liberalised early will be best placed to pursue an accelerated approach to switching off GSM, as part of a UMTS900 refarming strategy. Where refarming is already permitted or regulatory approval is imminent, there may be a good business case for a GSM operator to deploy UMTS900, in order to provide 3G rural coverage, increase 3G capacity, or both. Some operators, such as Elisa and DNA in Finland, have already deployed UMTS900 in rural areas and plan to create nationwide UMTS900 networks as quickly as possible. Once their nationwide 3G (UMTS900/2100) coverage has been completed, these operators will roll out UMTS900 in urban areas, using UMTS900 as an improved coverage layer and UMTS2100 as a capacity layer. Both operators intend to decommission their GSM networks as soon as possible, with Elisa aspiring to do so by 2015. Operators considering the business case for UMTS900 should therefore carefully assess whether a wide-scale roll-out could also support an earlier switch-off of the GSM network. However, where refarming measures are delayed or take time to implement, it becomes increasingly likely that an operator will migrate to LTE900, rather than UMTS900, which will in turn diminish the likelihood of early GSM switch-off being achievable.
While UMTS900 potentially opens the way for operators to bring forward the date when they switch off their GSM networks, there are several significant inhibitors.
- Migration from 2G to 3G handsets. Support for UMTS900 is rapidly becoming standard in handsets destined for EMEA and the Asia–Pacific region. However, operators launching UMTS900 networks are expecting the greatest usage initially to derive from mobile broadband services based on USB modems or datacards, and envisage that it will take longer for subscribers to replace their 2G handsets with phones supporting UMTS900/2100, especially in the current economic climate. If it is a strategic requirement to switch off the GSM network early, then an operator will need to incentivise its subscribers to upgrade their handsets more rapidly by subsidising 3G phones, but the costs and benefits of doing so would require careful consideration.
- GSM inbound roaming revenue. Even if an operator can quickly migrate its own subscribers to 3G handsets, it may decide to delay switching off GSM in order to provide continued support for roaming GSM visitors and to preserve roaming revenue. Operators will need to establish the tipping point at which the financial benefits of switching off the GSM network outweigh the residual roaming revenue. Should they decide to retain the GSM network to support roaming users, then costs could be minimised by outsourcing the management of the legacy network.
- Support for legacy M2M wireless communications. GPRS/EDGE technology currently supports a wide range of M2M applications, including security systems, point-of-sales transactions, vending machine restocking, remote metering, traffic monitoring, vehicle tracking and digital signage. There will be a requirement for high-bandwidth M2M applications to migrate to higher-speed technologies, such as 3G, WiFi or WiMAX, but for other applications this may never be necessary. Operators will therefore need to decide for how long they are prepared to continue supporting legacy M2M services.
Another issue influencing the timing of GSM switch-off will be the future use of any 1800MHz spectrum that an operator holds. Unlike UMTS900, for which the supply of network equipment and devices is rapidly increasing, there is relatively little vendor backing for UMTS1800 equipment or devices at present, and we regard refarming to LTE (or an alternative future technology) at 1800MHz as a more probable outcome than deploying UMTS1800. Operators are therefore likely to retain GSM services at 1800MHz until such point as refarming at that frequency is required for capacity reasons, and will deploy the most spectrally efficient technology at their disposal at that time. How quickly migration to LTE1800 becomes feasible will depend on the availability of the LTE standards, network equipment and devices. Vendors are reporting a strong operator push for LTE1800, and we expect the ecosystem to be viable within the next three years.
While mobile operators in many developed markets may opt to retain their GSM networks until the latter years of the next decade for the reasons discussed above, a rapid UMTS900 roll-out strategy may in some circumstances permit operators to adopt a more aggressive approach towards decommissioning their legacy network. However, any delays in either regulating or implementing spectrum refarming will increasingly favour operators prolonging the life of their GSM networks and refarming their 900MHz spectrum to LTE or another future technology rather than to UMTS.
Further discussion of this and related topics is available in Analysys Mason’s Wireless Networks research programme. For more information, please call our sales team on +44 1223 460600 or email research@analysysmason.com.
