Dr. Kim Kyllesbech Larsen, Senior Vice President of Technology Services for T-Mobile International speaks exclusively to IQPC Telecom IQ on how emerging markets can fully utilize 3G and amongst others, the risks and benefits operators face by skipping 3G and pushing forward with LTE.
Dr. Larsen has more than 10 years experience in the telecommunication industry, with national, as well as international, management roles within the T-Mobile organization. Dr. Larsen is heavily involved in business development in Growth Markets (e.g., China, India and South-East Asia in general) focusing on technology deployment, sourcing models and business modeling.
IQPC:
Could you please touch on and elaborate on what are some of the challenges when it comes to spectrum limitations for migrating to LTE?
Kim Larsen:
The key issues are suitable spectrum availability, backhaul transmission capacity, managing the customer-usage and the technology migration dynamics.
Let’s first define the minimum spectrum requirements for a technology migration. An operator would need sufficient spectral capacity (i.e., bandwidth), either un-occupied or under-utilized, to commence the migration from a legacy technology (e.g., 2G, 3G or WiMax) to LTE. You will often hear industry experts talking about deploying LTE at 2×20 MHz. This is of course a noble intention but would in many instances require a new spectrum to be acquired or the old spectrum to be freed-up and perfected.
A good example is a GSM-UMTS mobile operator with both 1.8 GHz and 2.1 GHz where the customer base and usage is migrating away from the 1.8 GHz band onto the 2.1 GHz band and in theory (at least) allowing capacity to be freed up in the 1.8 GHz for LTE. Whether sufficient capacity can be re-farmed will depend on factors such as spectrum utilization, the pace of migration (i.e., GSM -> UMTS) and growth of usage in the existing bands. An alternative would be for the operator to acquire (if available) FDD spectrum in the 2.5 GHz band or in the digital dividend band below 900 MHz. Yet another alternative is to acquire TDD based spectrum though such spectrum typically appears to another business model than most pure mobile players are used to.
Thus, one of the first challenges for an operator is to obtain frequency spectrum required for the migration to LTE. Such spectrum can come from the operator’s own portfolio or from new spectrum acquisition. If suitable frequencies and bandwidth are available, it becomes a question of possible regulatory approval (pending the spectrum rights associated with the existing spectrum) and the timing and management of the technology migration it self. If the operator is spectrum-limited new frequencies and bandwidth for the LTE migration would be required (e.g., by auction, spectrum-trading, beauty contests, partnerships, consolidation) It is almost always a less complex operation to use a new spectrum for technology migration rather than own an existing spectrum that would need to be re-farmed. However acquiring new spectrum can be costly and should be weighted against the possibility and costrequirements of re-perfecting its own spectrum. Particular for a Greenfield or a relatively new
financed mobile operation financing additional spectrum might not be a viable option. The good thing about LTE is that it’s been designed to be very versatile in terms of available bandwidth and frequencies required to commence operation. An operator, without new spectrum, could in principle start the LTE deployment with 2×1.25 MHz and as spectrum becomes available and perfected, extend the capacity to 2×20 MHz or whatever the operators limit might be (and of course eventually beyond as LTE evolve and spectrum becomes available). Furthermore, the most modernized 3G equipment promises a smoother transition from HSPA(+) towards LTE as a substantial part of the radio hardware can be re-used or partitioned for LTE.
In many emerging growth markets in Asia we have seen the launch of broadband wireless access (BWA) operations which have focused their business model on WiMax (i.e., TDD at 2.3 GHz, 2.6 GHz,..), such as the very successful Packet-1 Malaysia WiMax operation. These operations are almost all exclusively 1 spectrum-band wireless broadband operations with limited spectrum resources typically 1×20 or 1×30 MHz. Those newly started BWA business models will face significant obstacles to migrate to TDD-based LTE unless provided with additional spectrum for such migration. Moreover the significant differences in the underlying business model between WiMax and LTE, as well as the fundamental difference in core network and security-privacy architecture between WiMax and LTE, could make migration from WiMax to LTE complex and costly.
Thus, spectrum availability and managing the customer-usage and technology migration dynamics (i.e., GSM -> UMTS/HSPA -> LTE, GSM -> LTE …) are key challenges for a mobile operator wanting (or needing) LTE. Ignoring the economical challenges, newly established WiMax operations in the Asia-Pacific region would face high complexity and cost transitioning from WiMax to LTE. Operators with 2 or more frequency bands, i.e., typically the legacy Telco’s, will in general have less complexity issues with migration to LTE.
Last but not least ensuring adequate backhaul transmission capacity to the LTE site is another substantial challenge. The throughput being unleashed by LTE is orders of magnitude higher than what most mobile operators have been used to. Conventional means of backhaul, such as micro wave radios and leased lines, will not be able to scale (economically as well as operationally) with the increased nodal demand for throughput.
IQPC:
Which countries can be used as models when it comes to upgrading strategies from HSPA+ to LTE and what could these countries have done better in order to minimize cost while at the same time not compromising quality?
Kim Larsen:
Mobile operators with several years of 3G operations are likely to be in the process of, or have recently been, upgrading and modernizing their 3G radio infrastructure, enabling for example HSPA+ and IP backhauling. The modernization process has resulted in more cost efficient operations (i.e., energy and floor space) and significant higher quality compared to the first generation 3G infrastructure. The very latest 3G radio equipment supporting HSPA+ is in many cases expected to offer a relative smooth migration from HSPA+ to LTE (provided a compatible spectrum is available and can be perfected for LTE).
In order to better appreciate the connection between quality and economics it is important to understand that some mobile operations have based their 3G network (operating at 2.1 GHz) on their legacy 2G 900MHz grid. Though clearly more economical, such 3G networks are more likely to suffer capacity related quality issues than 3G networks overlaid a legacy 1800 MHz network. While LTE is expected to boost capacity and quality even for the sub-optimal 3G networks, it will critically depend on the carrier frequency how much improvement will be reached. In the end physics (i.e., link-budget) will determine the step in improvement we will see coming from deploying LTE. The quality and cost advantage will critically depend on the frequency range and the existing legacy networks being used as “backbone” for the LTE deployment. As long as several technologies are operated in parallel I would not expect to see much improvement in the absolute level of operational cost unless radical measures are taken.
Site and network infrastructure sharing could possible be the mean to radically reduce the LTE deployment capital cost (i.e., Capex avoidance) and achieve long-term (i.e., 5+ years) structural operational cost savings (and avoidance). One could imagine pooling of spectrum resources between operators, including new entrants with new spectrum that would allow carrier aggregation across various frequency bands (i.e., 700 MHz, 900 MHz, 1.8 GHz, all the way up to 2.6 GHz). Given the substantial re-design required to optimize networks for multiband carrier aggregation substantial Capex might be avoided by consolidating the LTE network deployment effort. Another possibility to minimize the LTE rollout cost is to make use of tower companies particular in emerging growth markets (i.e., such as India) that will provide additional Capex reduction measures for the LTE deployment offering multi-standard, multi-frequency base stations with fiber backhaul and possible managed services. Such scenarios are providing very clear up-front benefits. Though using tower companies often comes at expense of higher Opex, making it a very interesting option for a Greenfield or otherwise cash-restrained venture.
IQPC:
There are 124 countries where 3G has been launched but very few countries outside Europe and US have decent 3G coverage, how can emerging markets fully utilize 3G and will we see cases where countries skip 3G and focus on LTE?
Kim Larsen:
It depends on the market, maturity of 3G and spectrum availability (i.e., frequency and bandwidth). I would expect India to use the scarce 3G spectrum being auctioned this year (i.e., 2010) for badly needed voice capacity as the amount of 3G spectrum capacity is not really sufficient for mobile broadband data services (in my opinion). Thailand could in principle skip 3G and focus on LTE. Other countries in the Asia-Pacific region with low 3G demand, and without extensive financial commitment to rollout 3G, might as well benefit from a partition spare spectrum towards LTE, if allowed by the relevant regulator. However, the caveat is that LTE has been developed to deliver mobile broadband and not per se foreseen to deliver high-quality voice (at least initially). For an operator with existing and sufficient GSM capacity and quality, adding LTE rather than 3G could be an attractive alternative, allowing such an operator to provide (European/US-like) fixed-like broadband to households and consumers.
IQPC:
What sort of risks or benefits are countries and especially companies taking by skipping 3G and pushing forward with LTE?
Kim Larsen:
In emerging-growth markets, with relatively poor fixed broadband infrastructures, the LTE network could significantly boost the internet penetration of the general population with expected macro- as well as micro-economical benefits connecting the unconnected. It is faster and less costly to deploy a mobile broadband network than for example an extensive fiber-based network or modernizing a poor copper infrastructure. LTE would be far better at addressing such broadband demands in an emerging growth market than 3G (i.e., HSPA+), both from a capacity as well as quality point of view.
If a company believes that they eventually would require LTE within the next 2 to 5 years, economical and operational efficiency would be gained moving directly to LTE rather than first deploy 3G and then later migrate to LTE. However, spectrum availability, regulatory and competitive environment would need to be considered.
Early LTE deployment (i.e., 2010 – 2012) would have to depend on first-generation infrastructure and a relatively poorly developed device ecosystem (i.e., limited to dongles, data-cards and CPE-like equipment). In addition, the availability and quality of voice and SMS services on the early LTE solutions remain uncertain. This could limit the LTE business model possibilities in the early years compared to a full 3G mobile operation with mature infrastructure and device ecosystems. Choosing a HSPA+ to LTE migration path might be a less risky deployment strategy for many operators in emerging growth markets.
However, I do expect to see very innovative deployment scenarios and business models arising from emerging growth markets. Imagine the low-thrill operator that launches cheap voice-terminals supporting only GSM and LTE, avoiding the relative costly UMTS IPR overhead cost. Such an operator could in principle off-load prepaid voice towards VoIP over LTE, keeping the GSM capacity for high-quality high ARPU customers. Of course the emerging market WiMax business model (ala Packet-1 Malaysia) can easily be duplicated with LTE as well, focusing on providing high-quality (speed and availability) internet access to house holds and nomadic users in markets with poor fixed broadband alternatives (i.e., most emerging growth markets).
IQPC:
How can existing networks in emerging markets successfully implement LTE without putting too much strain on the current network and at the same time keeping costs at a minimal?
Kim Larsen:
For an existing mobile operator it will depend on the available bandwidth, suitable frequencies, and the degree of synergy that can be achieved with existing legacy networks (e.g., GSM, UMTS). Many emerging growth markets are having very dense areas and already experiencing substantial spectrum exhaustion within the existing GSM and UMTS frequency bands. Thus the likelihood of re-farming and freeing up legacy spectrum for LTE appears low. New spectrum would be required. New spectrum often means a different frequency range than the legacy spectrum. Thus, it is likely that some operators will experience a substantial link-budget mismatch between the legacy radio network and the optimum LTE network. This is pretty much the same situation that GSM 900 MHz operators have been facing with UMTS 2100 MHz.
I believe that LTE will be deployed by existing 3G/HSPA operators to alleviate cost and higher capital pressure from otherwise substantial 3G/HSPA capacity extension. The LTE technology is not only providing better broadband performance but also improved range (i.e., link-budget) that would mitigate deploying at a higher frequency range than the legacy 2G/3G networks.
Dr. Kim Larsen will be speaking on the panel at the 5th Annual Mobile Networks Evolution conference. For more information on the conference or to get the latest conference agenda please feel free to email the Marketing Manager Indre Riley at [email protected] or visit www.mobilenetworksasia.com