Академический Документы
Профессиональный Документы
Культура Документы
We predict that mobile traffic will grow ten-fold in developed countries by 2015. Mobile operators must meet this
demand while maintaining a healthy profit margin.
The challenge facing mobile operators in developed and developing markets is to reduce costs while profitably
meeting the rising demand for data. It is well recognised that this requires an all-IP network that is optimised to
carry data in both the access and core network.
3GPP LTE (Long Term Evolution) will be the likely choice for operators in developed markets, as well as many
emerging markets. As the latest evolution of the UMTS family, it is the natural successor to UMTS for GSM and
UMTS operators. However, deploying a new access and core network, while existing UMTS assets are not fully
utilised is difficult for many operators to accept. The predicted growth in data traffic means that most operators
must now recognise that there is a need to deploy LTE. So the question is not why, but where and when.
This paper considers the issues that are influencing operators decisions about where and when to deploy LTE.
The following questions are discussed:
The growing demand for data what are the main features and implications for RAN architecture?
What are the operational benefits of deploying LTE?
Before deploying LTE, is there room for growth in existing GSM/UMTS assets?
Will femtocells affect the deployment of LTE?
How will WiMAX, Wi-Fi and ADSL affect LTE deployment plans?
How does the availability of spectrum affect operator deployment strategies?
Finally, we conclude with our vision of deployment strategies that operators might employ.
The growing demand for data what are the main features and
implications for RAN architecture?
Compared with today, total wireless network traffic generated from voice and data services will increase ten-fold
by 2015 in developed regions.1 Traffic per cellular user per month in developed regions will rise from an average
of 56MB in 2008 to 455MB in 2015, as shown in Table 1. By 2015, 94% of traffic will be data and 74% of this
will be generated indoors. However, if operators continue to apply their current pricing models, the revenue per
megabyte will fall from USD0.86 in 2008 to USD0.12 in 2015. Developing regions will follow a similar trend.
Table 1: Key forecasts for mobile traffic in developed and developing regions [Source: Analysys Mason, 2009] 2
Traffic metric
2008 (developed)
2015 (developed)
2008 (developing)
2015 (developing)
57PB
557PB
50PB
307PB
56MB
455MB
22MB
83MB
Percentage of data in
total traffic
49%
94%
7%
79%
74%
34%
62%
It is envisaged that an LTE network will deliver against the rising demand for data at a profit. As a bonus, there
are operational benefits of deploying LTE.
deploying fewer access nodes compared with 3G UMTS. This is known as the flat-RAN architecture it
makes deployment more cost effective and easily scalable, although it should be noted that flat-RAN
architectures are also available with HSPA
We report here our baseline figures. In such a long-term prediction, a number of factors could result in traffic levels differing substantially
from our baseline forecast. In a low-traffic scenario, traffic per customer in developed regions is 4 times higher in 2015 than it was in 2008,
while in a high-traffic scenario it is 22 times higher.
For more information, see Heath, M. and Brydon, A., Wireless network traffic 20082015: forecasts and analysis, Analysys Mason
(Cambridge, 2008).
January 2009
allowing the early switch-off of GSM. Operators will be able to streamline their operations by switching off
GSM and replacing it with LTE. Operators must weigh the benefits of this strategy against the cost of
migrating subscribers, and the loss of roaming revenue and legacy services
offering an opportunity for RAN sharing. RAN share has often failed in the past because operators could not
agree on the value of existing assets. This will not affect LTE RAN sharing, because LTE is a new
deployment. RAN sharing is a very significant way of reducing the deployment cost of LTE.
LTE offers some operational savings and a significant increase in network capacity, but it is not the only
solution. When considering how to meet the demand for data, network operators have a number of possible
technology alternatives that can be mixed to create a cost-effective solution with a phased deployment,
ultimately leading to LTE. These alternatives include: GSM, UMTS, HSPA, HSPA+, femtocells and Wi-Fi.
Before deploying LTE, is there room for growth in existing GSM and
UMTS assets?
When considering whether or not to deploy LTE, operators are most likely to ask the following three questions:
Why deploy LTE to replace UMTS when we still have so many GSM subscribers?
Many countries have not yet deployed UMTS, why deploy LTE?
Is there the potential to develop UMTS before deploying LTE?
By the end of 2008, 3.5 billion mobile subscribers were using networks based on the GSM family of standards,
of which only 350 million subscribers were on 3G UMTS. In theory, operators could migrate a substantial
number of subscribers to UMTS before moving on to LTE. However, this ignores local regulatory constraints,
availability of suitable spectrum, local legacy services and the relatively limited roaming capability of UMTS.
Moreover, many GSM subscribers will not be prepared to pay for new handsets and/or data devices, nor is it
commercially viable for the operator to migrate them.
UMTS could be deployed more widely, because 3G spectrum has not yet been fully allocated in many countries,
but many wonder why deploy an older technology when LTE can be deployed for a similar cost.
In developed countries, many operators with UMTS HSPA networks have to decide whether to upgrade to
HSPA+ or to move directly to LTE. Most operators say they will carry out software upgrades to HSPA and
HSPA+ without deploying MIMO. Leading operators have, however, conducted public trials of HSPA+ with
hardware MIMO: clearly there is a feeling that HSPA+ still has growth potential. Some operators are
considering combining HSPA+ with a direct solution to the problem of indoor coverage, like femtocells or
Wi-Fi, as an effective data-optimised solution instead of LTE.
January 2009
business by a trained installation team provided by the operator. Backhaul is via the existing business external
connectivity.
The wide-scale use of femtocells may ease the pressure on operators to deploy LTE by transferring some of the
traffic from the macrocell network to the femtocells. Figure 1 shows the growth in traffic from 2008 to 2012 for
a mobile operator with 10 million subscribers. Demand will outstrip HSPA capacity in 20102011 and HSPA+
capacity by 2012. However, if indoor traffic is offloaded to a femtocell network, HSPA and HSPA+ will offer
sufficient capacity for 2012 and beyond.
Figure 1: Network traffic generated by an example mobile service mix, split between indoor and outdoor usage,
10000
LTE (20MHz)
9000
8000
7000
Network
capacities
6000
5000
4000
LTE (10MHz)
3000
HSPA+ (10MHz)
2000
1000
HSPA (10MHz)
0
2008
2009
Outdoor
2010
Indoor
2011
2012
While the femtocell solution is technically elegant, it is not straightforward to implement and mobile operators
must carefully consider their position before relying on a femtocell solution to meet their rising capacity demand.
Femtocells are not the only means of providing indoor coverage. Other options include Wi-Fi, combined with
WiMAX and ADSL.
For more information, see Heath, M. and Brydon, A., 3G Network Evolution from 2007 to 2012: HSPA+, LTE, WiMAX and femtocells,
Analysys Mason (Cambridge, 2008)
January 2009
January 2009
Spectrum and local regulatory constraints: the availability of spectrum and bandwidths, and the timing of
auctions. There may also be regulatory conditions attached to licences, such as population coverage
requirements.
The operators business objectives: for example early stage coverage driven or later stage capacity driven.
There are as many pathways to a data-optimised radio access networks as there are operators. In order to
summarise our discussion, we consider the three main operator types: the combined GSM and UMTS operator,
the UMTS-only operator and the GSM-only operator.
With direct indoor coverage using femtocells/Wi-Fi, the quality of coverage is likely to be far better than
macrocell coverage. Wi-Fi has the advantage of being a proven technology that can be deployed today with a
proven business model. Femtocells are a natural evolution for the mobile operator from a GSM/UMTS family
background, but the business case and technical challenges may yet prevent the wide-scale deployment of
femtocells. In both cases, success depends on the ADSL backhaul. This may make the mobile operator
unacceptably vulnerable to the fixed operator, especially given that our predictions are that the mobile operator
may take more than 20% of the fixed operators broadband business. A converged operator, with both fixed and
mobile networks, of course would not be concerned.
If it transpires that it is impossible to reconcile a profitable femtocell business case, and/or the operator does not
wish to rely on ADSL backhaul to provide its mobile broadband offering, then the operator will need to provide
January 2009
indoor coverage with the macrocell network. This is best done using low frequencies like 750MHz4 or 900MHz.
Whether the operator decides to buy 750MHz spectrum when it becomes available or to replace 900MHz GSM
with data-optimised access technology, such as LTE or HSPA+, depends on when the spectrum becomes
available, and how strong the demand for data is predicted to be. We believe that if the operator goes for the
refarmed 900MHz option then the operational costs of revisiting sites to make hardware upgrades mean that
most operators will deploy LTE in 900MHz when it becomes available rather than follow an upgrade path to
LTE via UMTS and HSPA.
It is also likely that the 2G/3G operator will take advantage of incidental indoor coverage from 2100MHz, in the
first instance. However, extending this strategy to include indoor coverage at 2600MHz may require a
prohibitively high site density. The site density required to achieve good coverage indoors at 2600MHz will need
to be weighed against alternative indoor coverage techniques like direct indoor coverage, refarmed 900MHz or
new, low-frequency spectrum.
GSM-only operators
Before LTE, the upgrade path for GSM-only operators would typically be through UMTS and on to HSPA.
Many are now considering moving directly to LTE, assuming that local spectrum regulations allow this strategy.
It is also possible that some operators in developing countries will consider using WiMAX, due to local
spectrum constraints.
The GSM-only operator must carefully consider how to provide indoor coverage. As with all other operators, the
choices are between direct indoor coverage solutions, like femtocells or Wi-Fi, or coverage from the macrocell
network, probably using low frequencies, like 900MHz. In developing countries, it is more likely that 900MHz
GSM will be retained because of its good rural coverage and the high proportion of voice usage. So a direct
indoor solution may be more appropriate, but this must be traded against the cost and potential revenue. An
alternative may be GSM 900 with Evolved EDGE, to provide the required data.
To be made available to mobile operators as a consequence of the switch over from analogue broadcast TV to digital in the UHF band.
January 2009
Summary
Because of the advantages it offers over existing technologies, it is a given that LTE will find worldwide use.
However, given the complexity of the decision process it is not surprising that many operators are adopting a
wait and see approach when it comes to upgrading their networks. It is predicted that LTE will be deployed by
early adopter operators in developed countries in 2009 and it will be in commercial use in 20102011. LTE
deployment will predominantly take place during the following two to three years.
There are many alternative paths to LTE. We have talked to numerous operators, and most say that they will not
add MIMO technology to HSPA, but will carry out software upgrades only, and yet operators are conducting
public trials of HSPA+ with MIMO technology. The availability of direct indoor coverage solutions like Wi-Fi
or femtocells will have a marked impact on the use of HSPA+. We expect that in developing countries
particularly, operators will consider deploying LTE, instead of UMTS/HSPA, as a replacement for GSM. In the
end, the path an operator chooses depends on four key factors: its existing technology, the local market
conditions, local spectrum availability and its business objectives. These are all very localised decisions. If
operators deliberate carefully, weighing up the costs and benefits of the various technology mixes and choosing
when to deploy rather than following the leader, they will profit from the growth in the demand for data. Any
mistakes made now, however, could cost them dearly in five to ten years time.
January 2009