Choose LTE2600 or LTE1800 from the position of the operator
What to put the mobile operator when choosing a frequency range for LTE technology?
The time has come when you can seriously think about the large-scale construction of LTE in Russia, this is not about a “test network” of 10% of 3G spots in cities with millions of people, but about full coverage. Check this hypothesis.
Let's imagine within the capabilities of the three main Russian mobile operators + Rostelecom / TELE2.
Who, in particular, makes a specific, non-conceptual, decision to begin deploying a new network? One of these people is the CEO of the operator. He also report to shareholders on the quality of return on investment. What input information should it take into account? One of the characteristics is the rate of return on investment. It will be largely determined by the existence of the market in the technology that is supposed to be mastered.
')
The graph shows the revenues received by Russian operators with the division by type of service (data from ABI research).
Data on the chart, both historical and forecast: 2013 - the last year, the statistics of which are taken into account. SMS stagnates, while the share of revenues from data transmission is growing (forecast). Revenues from voice communications are declining. We make a significant investment conclusion: data transfer becomes more important.
But from here it is not obvious how to ensure the “collection” of packet traffic from subscribers. For this, there are 2G (GSM900 / DCS1800), 3G 2100 MHz, LTE of many variations. We consider the technological and spectral mainstream, because it is necessary (for investors) to return investments, but not to thunder with the conclusion of a new technology and incur losses.
The next graph shows a rapid growth in the data per subscriber for Eastern Europe, including Russia. Such pictures like to bring manufacturers of telecom equipment, convincing the operator that it is time to introduce a new technology. But the previous chart cools the hotheads.
Intermediate conclusion: 2G technology is not likely to cope with the growth of data. It is necessary to make a decision on which technology is worth collecting packet traffic. The alternative is terrible: we do not develop a packet network (in the sense intended primarily for solving the problem of collecting packet traffic).
Let's see what else we know about the mobile market:
Intermediate conclusion: if the rate of return on investment is more important in a three-year term (well, it is more important considering the obsolescence of equipment), then we will solve the problem posed by the second schedule using 3G. At the same time, we mean that we will have to build one or two thousand 3G base stations (if they haven’t already been there) for the region (we limit ourselves to the scale of the region), in one year. It is possible for two years, but competitors do not wait and expand. Such a conclusion can be made more likely by an operator having a weak 3G coverage, i.e. for our example, TELE2 / Rostelecom or someone from the big three, who was late with 3G in some region, but began to doubt whether to invest in LTE right away, bypassing 3G.
We made a major decision on how to enter or expand technologies in a three-year term: the first phase of the plan is clear - 3G. What about LTE?
Managing a technological and progressive company (at least in its own advertising), our CEO may decide to spend part of the money on the “brand coverage” of the territory controlled by the LTE network, which will show foresight, progressiveness and form the basis for the second phase. And besides the return on investment, it will show investors a long-term development plan, materialized in the form of modern network fragments.
We confine ourselves to the top 5-10% of 3G coverage. Those. in the first phase, we will build not 1000 NodeB (3G base stations), but 950 (3G) + 50 (LTE). Then in a year, we will have a tolerable 3G coverage in the region and several LTE hot spots.
Let's choose which technology and LTE spectrum we will use. The choice is not very big, but responsible: TD-LTE and FDD. If there is a spectrum, we rely on the second one, but we understand that the occupied spectrum and the free one are not the same thing, it may be necessary to clear it (refarming).
In addition, do not forget to look at the schedule, how things are in the world. The graph below vertically shows the frequency of the spectrum in megahertz. Horizontal number of LTE networks in the specified range (~ 380 LTE networks analyzed, OVUM and GSMA data). It can be seen that equipment manufacturers and operators (ultimately phone manufacturers) came to the conclusion that there will be two main ranges for LTE (~ 140 networks).
We want to get a benefit, so we consider the 1800 and 2600 MHz frequency range for LTE FDD. Set priorities, because will have to order equipment that supports a specific range. Two upgrades instead of one is expensive, you need to make the right choice.
According to estimates by telecom equipment manufacturers and companies providing a radio network optimization service based on subscriber geolocation, up to 80% of traffic is generated on premises. This information clearly gives an advantage to the range of 1800 over 2600 MHz. Losses on penetration into the room for 1800 MHz are significantly lower than for the 2600 range. Traffic will collect better that range whose penetration is better. The difference between the WCDMA2100 and DCS1800 is quite noticeable, but at the expense of on average more sensitive UMTS terminals, the difference is leveled. But when comparing 2600 and 1800 MHz, other things being equal, we make a choice in favor of 1800.
Will subscribers of other networks (or, rather, networks of other countries) be able to use our LTE network? In addition to such an important administrative component as a roaming agreement between operators, in our case it is necessary to take into account whether the mobile terminal will support our standard and frequency range. We are interested in Band 3 - 1800 and Band 7 - 2600 MHz.
A selection of models of iPhone5s and Galaxy S4 - a slightly higher preference is given to the 1800 MHz range.
We make the following hypothesis: the frequency ranges supported by top terminals gradually “spread” across the whole range of sentences with some variations. So, probably, there will be some share of terminals supporting only 2600 or 1800, but not both at the same time. Based on how often the range is maintained in such models, we extrapolate the distribution to the entire terminal market and correlate it with the current distribution statistics in existing models. There is a great deal of attraction to the 1800 range, although the distribution of the networks does not fully explain this fact.
The simplicity of providing users of their network with terminals is not the same as ensuring the operation of roamers (statistical distribution of supported ranges by terminals). The Romer migration process is more random, so we take into account the actual situation on the market and can influence the capture of such roamers through the deployment of the range they prefer. And for its subscribers it is necessary only not to interfere with the use of selected devices. Conclusion: 1800 MHz - a little better.
The analysis done in isolation from the spectrum actually available to the operator, of course, requires an addition: if our operator is one of the Quartet, then it is very likely that he has 1800 MHz band used for DCS (2G). The case when 900 MHz is available and we plan to use it for LTE also does not exclude as a whole, but in our particular case, when we need support with a large number of subscriber devices (we plan to make money), we will not consider it yet.
So, we stop the choice for 1800 and here we understand that the entire available spectrum is occupied by DCS, which brings us money successfully, although not as successfully as before. A technical contradiction will have to be resolved: LTE technology in the 1800 range is needed to make money in the future and technology is not needed in this range in order to continue to bring money now (the gap is several years, this is not for us).
You can go the way of a compromise and make a partial release of the spectrum in the 1800 range from DCS (minimum, otherwise the capacity, and hence the traffic, drops significantly), but not for that we entered a contradiction to be content with a compromise. We will transfer the maximum of making money in 3G, as far as experience will allow us. Therefore, the maximum voice should go to 3G, it’s not for nothing that the construction of 950 BTS 3G is planned; The residual principle will allow subscribers in 3G 2100 MHz to use the Internet.
Now, the vacant capacity of the DCS 1800 can be started to be delivered under the LTE 1800. In practice, a compromise solution will most likely be obtained, since by the time of the beginning of refarming (we need a year for building 3G) packet traffic on 2G will grow and the capacity will not actually be released as much as expected, but, in any case, we have 3G, where you can redirect subscribers, trying to keep the quality on the reduced refarming DCS.
Choose the 1800 range. For LTE technology, the main thing is the availability of 3G. Outside of limitations, when you need to monitor quality or make money, the conclusion is different, but that's another story.
Distribution by region:
The time has come when you can seriously think about the large-scale construction of LTE in Russia, this is not about a “test network” of 10% of 3G spots in cities with millions of people, but about full coverage. Check this hypothesis.
Let's imagine within the capabilities of the three main Russian mobile operators + Rostelecom / TELE2.
Who, in particular, makes a specific, non-conceptual, decision to begin deploying a new network? One of these people is the CEO of the operator. He also report to shareholders on the quality of return on investment. What input information should it take into account? One of the characteristics is the rate of return on investment. It will be largely determined by the existence of the market in the technology that is supposed to be mastered.
')
Money
The graph shows the revenues received by Russian operators with the division by type of service (data from ABI research).
Data on the chart, both historical and forecast: 2013 - the last year, the statistics of which are taken into account. SMS stagnates, while the share of revenues from data transmission is growing (forecast). Revenues from voice communications are declining. We make a significant investment conclusion: data transfer becomes more important.
But from here it is not obvious how to ensure the “collection” of packet traffic from subscribers. For this, there are 2G (GSM900 / DCS1800), 3G 2100 MHz, LTE of many variations. We consider the technological and spectral mainstream, because it is necessary (for investors) to return investments, but not to thunder with the conclusion of a new technology and incur losses.
Technology
The next graph shows a rapid growth in the data per subscriber for Eastern Europe, including Russia. Such pictures like to bring manufacturers of telecom equipment, convincing the operator that it is time to introduce a new technology. But the previous chart cools the hotheads.
Intermediate conclusion: 2G technology is not likely to cope with the growth of data. It is necessary to make a decision on which technology is worth collecting packet traffic. The alternative is terrible: we do not develop a packet network (in the sense intended primarily for solving the problem of collecting packet traffic).
Let's see what else we know about the mobile market:
- 172% (161 active) - penetration of cellular communication in the market in relation to the population;
- 2.1 SIM-cards accounted for one subscriber (apparently, all SIM-cards are divided into subscribers, of which 108 million (unique, the numbers do not converge));
- 45% (2013 Q4) - 3G penetration;
- 2% - LTE penetration.
Intermediate conclusion: if the rate of return on investment is more important in a three-year term (well, it is more important considering the obsolescence of equipment), then we will solve the problem posed by the second schedule using 3G. At the same time, we mean that we will have to build one or two thousand 3G base stations (if they haven’t already been there) for the region (we limit ourselves to the scale of the region), in one year. It is possible for two years, but competitors do not wait and expand. Such a conclusion can be made more likely by an operator having a weak 3G coverage, i.e. for our example, TELE2 / Rostelecom or someone from the big three, who was late with 3G in some region, but began to doubt whether to invest in LTE right away, bypassing 3G.
Development plan
We made a major decision on how to enter or expand technologies in a three-year term: the first phase of the plan is clear - 3G. What about LTE?
Managing a technological and progressive company (at least in its own advertising), our CEO may decide to spend part of the money on the “brand coverage” of the territory controlled by the LTE network, which will show foresight, progressiveness and form the basis for the second phase. And besides the return on investment, it will show investors a long-term development plan, materialized in the form of modern network fragments.
We confine ourselves to the top 5-10% of 3G coverage. Those. in the first phase, we will build not 1000 NodeB (3G base stations), but 950 (3G) + 50 (LTE). Then in a year, we will have a tolerable 3G coverage in the region and several LTE hot spots.
We specify the plan
Let's choose which technology and LTE spectrum we will use. The choice is not very big, but responsible: TD-LTE and FDD. If there is a spectrum, we rely on the second one, but we understand that the occupied spectrum and the free one are not the same thing, it may be necessary to clear it (refarming).
In addition, do not forget to look at the schedule, how things are in the world. The graph below vertically shows the frequency of the spectrum in megahertz. Horizontal number of LTE networks in the specified range (~ 380 LTE networks analyzed, OVUM and GSMA data). It can be seen that equipment manufacturers and operators (ultimately phone manufacturers) came to the conclusion that there will be two main ranges for LTE (~ 140 networks).
We want to get a benefit, so we consider the 1800 and 2600 MHz frequency range for LTE FDD. Set priorities, because will have to order equipment that supports a specific range. Two upgrades instead of one is expensive, you need to make the right choice.
Signal penetration
According to estimates by telecom equipment manufacturers and companies providing a radio network optimization service based on subscriber geolocation, up to 80% of traffic is generated on premises. This information clearly gives an advantage to the range of 1800 over 2600 MHz. Losses on penetration into the room for 1800 MHz are significantly lower than for the 2600 range. Traffic will collect better that range whose penetration is better. The difference between the WCDMA2100 and DCS1800 is quite noticeable, but at the expense of on average more sensitive UMTS terminals, the difference is leveled. But when comparing 2600 and 1800 MHz, other things being equal, we make a choice in favor of 1800.
Roaming
Will subscribers of other networks (or, rather, networks of other countries) be able to use our LTE network? In addition to such an important administrative component as a roaming agreement between operators, in our case it is necessary to take into account whether the mobile terminal will support our standard and frequency range. We are interested in Band 3 - 1800 and Band 7 - 2600 MHz.
A selection of models of iPhone5s and Galaxy S4 - a slightly higher preference is given to the 1800 MHz range.
Model A1533 (GSM):
LTE (bands 1, 2, 3 , 4, 5, 8, 13, 17, 19, 20, 25)
Model A1533 (CDMA):
LTE (bands 1, 2, 3 , 4, 5, 8, 13, 17, 19, 20, 25)
Model A1453:
LTE (bands 1, 2, 3 , 4, 5, 8, 13, 17, 18, 19, 20, 25, 26)
Model A1457:
LTE (bands 1, 2, 3 , 5, 7 , 8, 20)
Model A1530:
FDD-LTE (1, 2, 3 , 5, 7 , 8, 20 ranges);
TD-LTE (bands 38 , 39, 40)
Galaxy S4 EU Model (GT-I9505)
LTE-TDD LTE-FDD
Band 1 - 2100
Band 3 - 1800
Band 5 - 850
Band 7 - 2600
Band 8 - 900
Band 20 - 800EU
We make the following hypothesis: the frequency ranges supported by top terminals gradually “spread” across the whole range of sentences with some variations. So, probably, there will be some share of terminals supporting only 2600 or 1800, but not both at the same time. Based on how often the range is maintained in such models, we extrapolate the distribution to the entire terminal market and correlate it with the current distribution statistics in existing models. There is a great deal of attraction to the 1800 range, although the distribution of the networks does not fully explain this fact.
User terminals
The simplicity of providing users of their network with terminals is not the same as ensuring the operation of roamers (statistical distribution of supported ranges by terminals). The Romer migration process is more random, so we take into account the actual situation on the market and can influence the capture of such roamers through the deployment of the range they prefer. And for its subscribers it is necessary only not to interfere with the use of selected devices. Conclusion: 1800 MHz - a little better.
Spectrum availability and refarming
The analysis done in isolation from the spectrum actually available to the operator, of course, requires an addition: if our operator is one of the Quartet, then it is very likely that he has 1800 MHz band used for DCS (2G). The case when 900 MHz is available and we plan to use it for LTE also does not exclude as a whole, but in our particular case, when we need support with a large number of subscriber devices (we plan to make money), we will not consider it yet.
So, we stop the choice for 1800 and here we understand that the entire available spectrum is occupied by DCS, which brings us money successfully, although not as successfully as before. A technical contradiction will have to be resolved: LTE technology in the 1800 range is needed to make money in the future and technology is not needed in this range in order to continue to bring money now (the gap is several years, this is not for us).
You can go the way of a compromise and make a partial release of the spectrum in the 1800 range from DCS (minimum, otherwise the capacity, and hence the traffic, drops significantly), but not for that we entered a contradiction to be content with a compromise. We will transfer the maximum of making money in 3G, as far as experience will allow us. Therefore, the maximum voice should go to 3G, it’s not for nothing that the construction of 950 BTS 3G is planned; The residual principle will allow subscribers in 3G 2100 MHz to use the Internet.
Now, the vacant capacity of the DCS 1800 can be started to be delivered under the LTE 1800. In practice, a compromise solution will most likely be obtained, since by the time of the beginning of refarming (we need a year for building 3G) packet traffic on 2G will grow and the capacity will not actually be released as much as expected, but, in any case, we have 3G, where you can redirect subscribers, trying to keep the quality on the reduced refarming DCS.
Conclusion
Choose the 1800 range. For LTE technology, the main thing is the availability of 3G. Outside of limitations, when you need to monitor quality or make money, the conclusion is different, but that's another story.
Some facts
(reports and experience)- According to a survey of 40+ operators, LTE input increased their ARPU by 10%.
- ARPU in many global reports counts on Connection, not user.
- The average consumption of a 1 GB package is 35%.
- The average monthly user traffic for HSPA (3G) networks is 1.1 GB, and after entering LTE, these operators report an average monthly user traffic of 2.1 GB.
Distribution by region:
Source: https://habr.com/ru/post/219317/
All Articles