The case for Wi-Fi offload the costs and

Скачать презентацию The case for Wi-Fi offload the costs and

61be38f5e60d4d6a0f29d0ca215dca09.ppt

Количество слайдов: 58

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Research Report The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks December 2011 Terry Norman and Richard Linton © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 2 Contents Slide no. 5. 6. 7. 8. 9. 10. 11. 12. 13. 22. Indoor Wi-Fi only model 23. The total deployment costs for an indoor Wi-Fi access point are EUR 480 and total recurring costs are EUR 560 per annum 24. The total cost of providing additional capacity using the indoor Wi-Fi only model is EUR 124 910 25. Outdoor plus indoor Wi-Fi model 26. Our outdoor plus indoor Wi-Fi model assumes a mesh architecture 27. The total deployment cost for an outdoor plus indoor Wi-Fi mesh network is EUR 61 850 and total recurring costs are EUR 22 900 per annum 28. The total cost of providing additional capacity using first outdoor, then indoor Wi-Fi cells is EUR 260 317 29. The cost of capacity required to relieve congestion in a 3 G macro-cell is lowest for indoor Wi-Fi only 30. The leading operators’ propositions 31. In the near future, next-generation Wi-Fi hotspots will be an important strategic element in MNOs’ churn reduction 32. The Wi-Fi hotspot value chain encompasses a number of commercial relationships 33. Commercial entities, such as i. Pass, that aggregate public Wi-Fi hotspots are at the top of the value chain 34. National, local and regional operators of Wi-Fi hotspots 35. Operators may form partnerships to expand their hotspot footprints, but should complement this with their own deployments 14. 15. 16. 17. 18. 19. 20. 21. Executive summary [1] Executive summary [2] Executive summary [3] Recommendations [1] Recommendations [2] The drivers of Wi-Fi offload Growth in mobile traffic and number of connections will be substantial up to 2016 Traffic growth and falling revenue are the principal drivers of Wi-Fi offload To meet traffic demand through macro network expansion would be prohibitively expensive Mobile network operators will need to reduce network carriage costs by 50% Operators are seeking ways to reduce costs and Wi-Fi is one such way The cost of deploying commercial Wi-Fi cells We present an analysis of coverage costs in three scenarios The LTE-only model To provide additional capacity using 4 G only, a total of more than EUR 600 000 may be spent on a single, congested site © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Contents Slide no. 36. In the UK, only one MNO has so far announced plans to deploy its own Wi-Fi network 37. Elsewhere, a number of large-scale Wi-Fi roll-outs have recently been announced 38. Recent developments support the claim that Wi-Fi is on the verge of becoming an integral network component for MNOs 39. Operators worldwide are busy with hotspot deployments [1] 40. Operators worldwide are busy with hotspot deployments [2] 41. Customers have come to expect that certain commercial premises will have Wi-Fi access points and this helps to stimulate the market 42. The emergence of NGH signals the end of Wi-Fi’s role as a complementary component of MNOs’ service offering 43. Alternatives to, and problems with, Wi-Fi offload 44. Wi-Fi is cheaper than LTE and operates in free spectrum, so is there a catch? 45. Connecting to a Wi-Fi hotspot should be as simple as connecting to a mobile network and problems may harm the user’s bond with the operator 46. The Wi-Fi industry must standardise in order to deliver carrier-grade products and services 47. Backhaul is still a challenge for all small-cell solutions, not only for Wi-Fi 48. MNOs will need the services of trusted partners to backhaul Wi-Fi 49. Annex: additional data 50. Indoor Wi-Fi only model assumptions: breakdown of total deployment and recurring costs 51. Outdoor Wi-Fi unit assumptions: breakdown of total deployment and recurring costs 52. Mesh networks in more detail [1] 53. Mesh networks in more detail [2] 54. About the authors and Analysys Mason 55. About the authors 56. About Analysys Mason 57. Research from Analysys Mason 58. Consulting from Analysys Mason © Analysys Mason Limited 2011 3

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 4 List of figures Figure 1: Comparison of cumulative RAN costs (opex + capex) for capacity upgrade investments Figure 2: Average mobile network traffic per month per connection, 2011– 2016 Figure 3: Average number of mobile network connections, 2011– 2016 Figure 4: Monthly traffic from mobile connections, 2011– 2016 Figure 5: Revenue per gigabyte of mobile broadband traffic, 2011– 2016 Figure 6: Annual capex on RAN, Western Europe, 2011– 2016 Figure 7: Annual capex in RAN assuming 50% reduction in network carriage costs, Western Europe, 2011– 2016 Figure 8: Capex for new LTE build Figure 9: Opex for new LTE build Figure 10: Cumulative capital investment (capex and opex) in LTE-only model Figure 11: Total deployment costs for an indoor Wi-Fi access point 1 Figure 12: Total recurring costs for an indoor Wi-Fi access point Figure 13: Cumulative capital investment (capex and opex) in indoor Wi-Fi only model Figure 14: Outdoor plus indoor Wi-Fi model architecture Figure 15: Total deployment costs for an outdoor plus indoor Wi-Fi mesh network © Analysys Mason Limited 2011 Figure 16: Total recurring costs for outdoor plus indoor Wi-Fi mesh network Figure 17: Cumulative capital investment (capex and opex) in outdoor plus indoor Wi-Fi model Figure 18: Comparison of cumulative costs for three deployment models Figure 19: The Wi-Fi hotspot value chain Figure 20: Number of access points accessible to the largest hotspot aggregators Figure 21 a: Examples of recent, large-scale deployments of Wi-Fi by MNOs [1] Figure 21 b: Examples of recent, large-scale deployments of Wi-Fi by MNOs [2] Figure 1 A: Total deployment costs for an indoor access point Figure 2 A: Total recurring costs for an indoor access point Figure 3 A: Total deployments costs for outdoor Wi-Fi network Figure 4 A: Total recurring costs for outdoor Wi-Fi network Figure 5 A: Outdoor Wi-Fi model architecture Figure 6 A: Illustrative average access point capacity versus number of access points

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Executive summary Recommendations The drivers of Wi-Fi offload The cost of deploying commercial Wi-Fi cells The leading operators’ propositions Alternatives to, and problems with, Wi-Fi offload Annex: additional data About the authors and Analysys Mason © Analysys Mason Limited 2011 5

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 6 Executive summary [1] § The demand for mobile data traffic and falling revenue per gigabyte are forcing operators to seek ways to reduce network carriage costs and to restore margins. We estimate that, on average, the global demand for mobile data traffic will grow at a CAGR of 52%. If operators endeavour to meet this demand using conventional macro site solutions, the cost of RAN equipment will increase eightfold. We argue that this is prohibitively expensive. In Western Europe alone, capex on RAN would increase from USD 5 billion in 2011 to approximately USD 40 billion in 2016. § In order to reduce capex on RAN to a manageable level, operators will need to improve the efficiency of traffic delivery and to reduce network carriage costs by 50%. In Western Europe, for example, taking these steps would cause capex on RAN to rise not to about USD 40 billion, but instead to about USD 10 billion in 2016, which would represent an acceptable growth in capex over the period 2011 to 2016. § One way to reduce network carriage costs, which is attracting a great deal of interest from MNOs, involves carrying a proportion of the traffic on a cost-efficient small cell. Because it is widely deployed and competitively priced, Wi-Fi is the leading candidate ‘small cell’ technology. § During September 2011, a number of large-scale Wi-Fi roll-outs were announced by MNOs. For example, KDDI in Japan (in conjunction with Ruckus Wireless) has plans for a total of 100 000 hotspots. Companies are also being set up to offer wholesale and aggregate Wi-Fi services. Towerstream is deploying 1000 routers in seven square miles of Manhattan, NY, USA, which is equivalent to 54 hotspots per square kilometre. § We have modelled the cost to build and operate mobile radio coverage over an urban area of 0. 8 km 2, which is typical of an urban 3 G site. We have compared the cost to upgrade the site in three different ways, each designed to support a CAGR of 52% in mobile data traffic. In the first scenario, we assumed that the 3 G site was upgraded to LTE and that further LTE sites were built as necessary to meet the traffic growth. In the second, indoor Wi-Fi access points are used to augment capacity so that total mobiledata traffic may grow at a rate of 52% per annum. In the third case, a mix of outdoor and indoor Wi-Fi access points are used to augment capacity to allow total mobile-data traffic to rise at a CAGR of 52%. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Executive summary [2] § The costs of indoor and outdoor Wi-Fi are both significantly lower than those of upgrading to 4 G. The final cumulative RAN costs (capex plus opex) are EUR 124 910 for the indoor Wi-Fi solution, EUR 260 317 for the indoor plus outdoor Wi-Fi solution and EUR 623 431 for the 4 G upgrade solution. § Costs of indoor Wi-Fi are the lowest because the equipment is much cheaper and there are many more users per square metre indoors. As a result, the traffic density per square metre is far higher than for either macro or outdoor Wi-Fi. § Not only are costs of indoor Wi-Fi significantly lower than those of outdoor and macro, but also they increase in a gradual fashion, which allows costs to be matched more closely to demand revenue, which is a further advantage of indoor Wi-Fi. § It is important to note that this report does not cover the use of Wi-Fi by domestic users in the home or the office, except where we refer to its use in the context of stimulating the market for the commercial use of Wi-Fi by MNOs. Where we refer to Wi-Fi, we mean the commercial use of Wi-Fi by MNOs to augment existing macro network capacity through hotspots provided by the MNO, either directly or through a third party. © Analysys Mason Limited 2011 Figure 1: Comparison of cumulative RAN costs (opex + capex) for capacity upgrade investments [Source: Analysys Mason, 2011] 7

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 8 Executive summary [3] § The mobile operator may find it very difficult to attach enough users to the outdoor Wi-Fi unit completely to relieve the congested macro cell that it is supporting. Almost 95% of tablet users and 70% of smartphone users will be found indoors, leaving very few heavy users of data outdoors. Wi-Fi carrier frequencies do not penetrate into buildings, so outdoor Wi-Fi units alone can at best carry only 10% of the heavy data traffic. This will not be sufficient to prevent the macro site becoming overloaded. It is likely that indoor Wi-Fi units will also be required. § Although outdoor Wi-Fi could theoretically be used as a cost-effective means of providing radio mobile data coverage to an entire macro cell area, could be reduced by deploying fewer access points and concentrating them on areas of micro-congestion within the cell. § The capacity limitations of copper backhaul may well constrain traffic growth on indoor Wi-Fi units within a very short time – one to two years. Wi-Fi is best deployed as an indoor solution, when fixed backhaul, using copper pairs, should be readily available. However, the capacity of fixed copper backhaul, typically less than 30 Mbps, will quickly be exceeded by the traffic throughput required of the Wi-Fi unit. § For indoor deployments, the operator’s challenge is to access and acquire the site. This may mean leasing access to the site from the site owner, or establishing a commercial relationship with the site owner, for example a chain of cafés. This can be a complex, difficult and time-consuming process. § The Wi-Fi industry must overcome significant issues in order to deliver ‘carrier-grade’ products and services. In order to drive global adoption of Wi-Fi by mobile operators, the areas of authentication, encryption and policy control must be standardised. Femtocells overcome these issues, but are not yet competitively priced. We expect the current battle between the technologies to be decided during the next two years. The final resolution will probably be the convergence of the two technologies. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 10 Recommendations [1] § Operators must reduce network carriage costs by 50% in order to support a CAGR of 50% in data traffic. There are various techniques at their disposal to achieve this. In particular: network sharing, small cell solutions, such as Wi-Fi, self-optimising networks and low-frequency spectrum, such as 800 MHz. Operators should investigate the costs and benefits of these major network carriage cost-reduction techniques. § Operators should trial Wi-Fi as a service delivery technology. Most operators are already trialling Wi-Fi as an offload technology, and should continue these trials, but extend them to consider the services that these Wi-Fi solutions can deliver, for example video. They should determine how this affects their strategies and business models. In terms of business models, operators should contrast and compare two fundamentally different models: B 2 B and B 2 C. Both have their advantages and disadvantages. § Operators should seek appropriate Wi-Fi partners. It is unlikely that MNOs will have the operational capability to deliver an effective Wi-Fi network. The partners should be able to deliver carrier-class Wi-Fi immediately. They should also be able to deliver Wi-Fi capability on the scale and density of access points that operators require. Few partners have this capability currently. § Pure mobile MNOs should seek trusted fixed partners. If the operator is ‘mobile only’, having no fixed sister company, it should establish partnerships for Wi-Fi backhaul with trusted fixed partners. A pure MNO Wi-Fi strategy is particularly vulnerable to the strategies of a fixed backhaul provider. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 11 Recommendations [2] § Companies that intend to establish an outdoor Wi-Fi footprint should carefully consider the available technology options. The cost–performance characteristics of mesh networks vary significantly. Some operators with existing footprints are refreshing the technology in their networks in order to deliver better streaming video services to subscribers. Wi-Fi vendors are lining up to provide a range of innovative products and services into the marketplace. § Fixed broadband service providers should consider providing managed Wi-Fi services. This strategy is supported by the experience of such companies in providing managed services to end users. Also, since both indoor and outdoor Wi-Fi deployments often require fixed backhaul connections, service providers can promote the use of their fixed broadband products. Note that fixed broadband service providers, especially those that do not offer cellular services, can use (fast) Wi-Fi to dampen the impact of 4 G services, when they arrive. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 13 Growth in mobile traffic and number of connections will be substantial up to 2016 § In developed markets, growth in mobile traffic is driven by the proliferation of smart devices (including mobile broadband devices, smartphones and tablets). 1 This is evidenced by a report released in August 2011 by Ofcom, the UK’s telecoms regulator, which states that the UK is “addicted to smartphones”. 2 Over 50% of adults, and nearly 50% of all teenagers, own a smartphone; 37% of adults and 60% of teens are “highly addicted” to smartphones, which are beginning to affect social behaviour. New services, particularly short videos and social networking, are likely to provoke continuing growth until 2016 and beyond. § In developed markets, the average monthly mobile traffic per connection will increase from 182 MB in 2011 to 892 MB in 2016, at a CAGR of 31%. Similarly in developing markets, mobile traffic per device will rise from 87 MB to 301 MB per month at a CAGR of 28%. The number of connections in emerging markets will grow substantially, driving up the global number of connections from 5 billion to 7 billion at a CAGR of 7%. Interestingly, penetration has almost reached saturation in developed markets and any rise in device numbers will be because of secondary device purchases. § The net effect is massive annual growth in mobile traffic. 1 See Analysys Mason’s Wireless network traffic worldwide: forecasts and analysis 2011– 2016. 2 Ofcom (London, UK, 2011), Communications Market Report: UK. Available at http: //stakeholders. ofcom. org. uk/binaries/research/cmr 11/UK_CMR_2011_FINAL. pdf © Analysys Mason Limited 2011 Figure 2: Average mobile network traffic per month per connection, 2011– 2016 [Source: Analysys Mason, 2011] Figure 3: Average number of mobile network connections, 2011– 2016 [Source: Analysys Mason, 2011]

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 14 Traffic growth and falling revenue are the principal drivers of Wi-Fi offload § We expect mobile data traffic to grow at a CAGR of 52% from 570 PB per month in 2011 to 3243 PB per month in 2016. 1 In 2016 in developed markets, traffic (voice and data) levels will be six times higher than in 2011; data-only traffic will seven times greater. Similarly in emerging markets, traffic levels are expected to be seven times higher in 2016; data traffic alone is forecast to be eight times greater. § However, if operators continue to use the flat-rate pricing model, we predict revenue of USD 4 per gigabyte by 2016 in both developed and emerging markets. Clearly, the extent to which tiered pricing is adopted in the future will affect this forecast. We are already observing moves away from flatrate pricing, as most operators are capping downloads. 1 See Analysys Mason’s Wireless network traffic worldwide: forecasts and analysis 2011– 2016. The 52% growth that we quote throughout this report describes mobile data. Currently, MNOs would deliver this traffic via the macro network. In our calculation of 52%, we account for the fact that a substantial portion of traffic being generated by the wireless device is carried over a Wi-Fi network. Since the mobile operator plays no real part in this, we class this as passive offload, or passive Wi-Fi offload. In future, the macro network may include carrier class Wi-Fi as part of the network delivery architecture. This we class as active Wi-Fi and, where appropriate, describe the process of transferring traffic from the conventional macro network onto this active Wi-Fi network as active Wi-Fi offload. © Analysys Mason Limited 2011 Figure 4: Monthly traffic from mobile connections, 2011– 2016 [Source: Analysys Mason, 2011] Figure 5: Revenue per gigabyte of mobile broadband traffic, 2011– 2016 [Source: Analysys Mason, 2011]

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks To meet traffic demand through macro network expansion would be prohibitively expensive § We have estimated the capex on RAN that would be necessary in Western Europe to service the growth predicted in our wireless traffic forecast, modelling two scenarios. § In the first scenario, we have assumed that operators continue to meet the traffic demand by using macro cell sites, enhancing HSPA sites to HSPA+ and deploying LTE from 2013. Operators would also deploy dual carrier technology on existing HSPA sites and deploy 2× 2 MIMO technology. The average cost of a base station was assumed to be USD 100 000, amortised over four years. Neither opex nor backhaul costs were included. LTE was assumed to be twice as efficient as HSPA+, in that each cell could support twice as many subscribers. § In the early years, the growth in capex is steady, but still doubles from USD 5 billion to 10 billion by 2013. After 2013, the impact of traffic growth is substantial. Despite being offset by the deployment of more-efficient LTE, capex rises substantially by 2016 to more than USD 40 billion. We judge this to be a figure far beyond the means of MNOs. As such, we predict that operators will be unable to build sufficient network capacity to support 52% per annum growth in mobile data traffic using only a macro base station strategy. © Analysys Mason Limited 2011 Figure 6: Annual capex on RAN, Western Europe, 2011– 2016 [Source: Analysys Mason, 2011] 15

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 16 Mobile network operators will need to reduce network carriage costs by 50% Figure 7: Annual capex in RAN assuming 50% reduction in network carriage costs, Western Europe, 2011– 2016 [Source: Analysys Mason, 2011] § In our second scenario, we have assumed that MNOs will achieve a 50% reduction in network carriage costs by 2016 – in other words, the cost of delivering 1 MB of data will be half that in 2011. § In this scenario, we see that the growth in capex in the RAN is gradual, doubling over five years from 2011 to 2016. § We argue that this is a more reasonable figure and that, if operators could achieve 50% reduction in network carriage costs over the next five years, it would be possible to support annual growth of approximately 52% in mobile traffic. § However, to achieve such a saving in capex is not a trivial matter. LTE alone is unlikely to offer this. Much of the cost efficiency of LTE is realised in opex savings, for example in power efficiency. This is at the cost of further capex, as it requires existing GSM and UMTS base stations to be replaced with multi-standard base stations that are more power efficient. § There are many ways in which operators can achieve such efficiency improvements, for example: by network sharing; by deploying low-frequency carriers, particularly in rural areas; and by bandwidth optimisation. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 17 Operators are seeking ways to reduce costs and Wi-Fi is one such way § Operators will employ a number of strategies to enhance their network capacities and to reduce network carriage costs (the cost of delivering each megabyte of data). § Wherever possible, operators will upgrade existing HSPA cells to HSPA+, increasing the peak download rate from 14. 4 Mbps to 21 Mbps. MIMO antenna enhancements may further increase this peak to 42 Mbps. Operators will also employ dual carrier technology for HSPA cells where possible, effecting a dual-carrier trunking gain. § From 2013, mass deployments of LTE are expected. LTE as a technology has better spectrum-utilisation properties than HSPA: the spectral efficiency per hertz of bandwidth is similar, but the bandwidth per cell can be twice or four times that of an HSPA cell. § Operators will employ low-frequency spectrum – typically digital dividend spectrum – in order to take advantage of its greater coverage per cell, compared with that of higher frequencies. Typically, the area of a cell at 800 MHz is three times that of a cell at 2. 1 GHz. § Operators will refarm existing spectrum (GSM 900) as it becomes available and instead deploy more-spectrally efficient HSPA+ or LTE. § Wherever possible, operators will seek to share networks. Passive networks can reduce capex by as much as 20– 30%. 1 § Operators will use other cost-saving techniques, such as self-optimising networks and capacity management and optimisation techniques. § Even with these cost-reduction techniques, operators are still facing a dilemma. Dramatic growth in traffic levels are not challenging for operators, but the combination of traffic growth and falling revenue per gigabyte is putting pressure on operators’ cash flow. § Unless operators find ways to reduce the cost of delivering traffic, they will be caught up in a spiral of decreasing profit as they roll out expensive networks to maintain a competitive network quality of service and so retain customers. Wi-Fi may be one means of reducing the cost of data delivery substantially. 2 1 See Analysys Mason’s Transform the economics of your wireless business with infrastructure sharing. 2 It is important to note that this report is not concerned with the use of Wi-Fi by domestic users in the home. Where it refers to Wi-Fi, this is the commercial use of Wi-Fi by MNOs to augment existing macro network capacity by providing Wi-Fi hotspots, either directly or through a third party. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 19 We present an analysis of coverage costs in three scenarios § In this section, we describe the cost of building and operating mobile radio coverage with and without the benefit of Wi-Fi over an urban area of 0. 8 km 2 (with a radius of 500 m), which is typical of an urban 3 G site. We assume that mobile services are provided to the area by a 3 G radio site, but that the radio site is congested and unable to support further traffic without being upgraded in some way. We compare the cost of upgrading the site in three different ways, each designed to support 52% growth per annum in mobile data traffic growth: - in the first scenario, we assume that the 3 G site is upgraded to LTE and that further LTE sites are built as necessary throughout the timeline of the model (2012– 2018) to meet the growth in traffic. No Wi-Fi is involved in this base line LTE-only model - in the second, the 3 G site is not upgraded so it remains fully utilised, and Wi-Fi access points are used to augment capacity in order to meet growth in total mobile data traffic of 52% per annum. The Wi-Fi access points are assumed to be located only indoors only in this indoor Wi-Fi only model - in the third case, the 3 G site is again not upgraded, but this time a mix of outdoor and indoor Wi-Fi access points are used to augment capacity in order to meet growth in total mobile data traffic of 52% per annum. This is the outdoor plus indoor Wi-Fi model. § In the following sections, we estimate and compare the cost of building and operating each of these upgrade solutions. § In all three scenarios, we have assumed that 95% of laptop use is indoors and 5% outdoors, while smartphone use is 30% outdoors and 70% indoors. We have not modelled the mobility of users through the day. § All equipment costs assume annual inflation of 3% and annual price reduction of 5% on equipment. The net effect is a reduction in equipment costs of 2% per annum. § There are several methods of accounting for the cost of capital, including DCF, NPV and WACC. Given that these are sensitive to the underlying assumptions made, we have instead presented our results as unamortised expenditure without cost of capital allowances. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks The LTE-only model § We first list the principal costs incurred when upgrading from 3 G to 4 G (LTE) and when building a new LTE site, then summarise the annualised (opex and amortised capex) cost of meeting the growth in mobile data traffic forecast. Figure 8: Capex for new LTE build [Source: Analysys Mason, 2011] § To upgrade the site from 3 G to 4 G (LTE), there are two principal costs: that of radio equipment, including backhaul, and that of spectrum. There is no additional opex, as we assume that site opex is accounted for in the 3 G site costs. The total cost to upgrade from 3 G to 4 G is EUR 30 000. § The spectrum cost is assumed to be 0. 15 EUR/MHz/pop, the population per site is assumed to be 9600, and the bandwidth is 20 MHz. The total cost of spectrum is EUR 28 800. This cost is apportioned to the area and not to an individual site, as it is derived from the population covered by the site. There is therefore a one-off apportionment of EUR 28 800. § For the cost of a newly constructed LTE site, the capex and opex are shown in Figures 8 and 9. The capex for a newbuild LTE site is EUR 110 000 and the opex is EUR 31 000. © Analysys Mason Limited 2011 Figure 9: Opex for new LTE build [Source: Analysys Mason, 2011] 20

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks To provide additional capacity using 4 G only, a total of more than EUR 600 000 may be spent on a single, congested site § As the demand for traffic rises (in 2012), initially the 3 G site is upgraded to LTE and further LTE sites are built as necessary to meet the predicted traffic growth. § The cost of this upgrade is EUR 30 000. The annual cost also includes the cost of spectrum for the area (EUR 7624). § In 2013, demand within the 0. 8 km 2 site area exceeds the capacity of the upgraded site and it is necessary for the operator to build a new site in the area. § The cost of this new site is EUR 110 000. The accumulated spend in 2013 rises to EUR 196 656 (for LTE upgrade, cost of spectrum, new site and annual opex). § The capacity of the site is sufficient to meet the customer demand for traffic until 2015, when it again becomes necessary for the operator to build a new site, at a cost of EUR 110 000. The total accumulated cost is EUR 401 272. § The capacity of the site is then sufficient to meet traffic demand to the end of the timeframe of the model, 2018. The final cumulative cost is EUR 623 431. § This cost curve serves as a benchmark against which the cost of building the two Wi-Fi solutions may be compared. © Analysys Mason Limited 2011 Figure 10: Cumulative capital investment (capex and opex) in LTE-only model [Source: Analysys Mason, 2011] 21

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 22 Indoor Wi-Fi only model § In the indoor Wi-Fi only model the fundamental assumption is again that mobile services are provided to the area by a 3 G radio site, but that the radio site is congested and is not able to support any further traffic. The operator’s solution is to build several indoor Wi-Fi sites. The number of indoor Wi-Fi sites is chosen to be just enough to ensure that the 3 G site is kept at the point of capacity saturation. § The purpose of this model is to understand how much this might cost the MNO, and to provide a comparison with an LTE-only alternative. § An indoor Wi-Fi-only model should be very cost effective. The capex and opex are low and indoor Wi-Fi should be able to carry a large proportion of the traffic load, since we assume that 95% of laptop users and 70% of smartphone use is indoors. § The model assumes that the operator has the ability to transfer indoor users onto the Wi-Fi network in preference to the congested 3 G network. § As with the LTE-only model, we first list the principle cost items incurred when upgrading from 3 G to 4 G (LTE) and when building a new LTE site. We then summarise the annualised cost (opex and amortised capex) of meeting the 52% per annum growth in mobile data traffic. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 23 The total deployment costs for an indoor Wi-Fi access point are EUR 480 and total recurring costs are EUR 560 per annum Figure 11: Total deployment costs for an indoor Wi-Fi access point 1 [Source: Analysys Mason, 2011] § The cost of deployment of an indoor Wi-Fi access point is much lower than that of outdoor access point (typically EUR 6185). Hardware costs contribute EUR 120, representing 20% of total deployment costs. Hardware costs are modest because indoor locations provide a shielded environment and access points are standalone, so quite basic technology may be deployed. § Backhaul costs contribute EUR 100, or 21% of total deployment costs. Note that service providers’ business models often require the owner of the premises to arrange and pay for his own backhaul. We have used list prices for backhaul circuits, but contract prices may be lower. Figure 12: Total recurring costs for an indoor Wi-Fi access point [Source: Analysys Mason, 2011] § Project management and installation contribute 14% and 10% of total deployment costs, respectively. § The recurring costs are dominated by those of backhaul, which are EUR 450 per annum (80% of the recurring costs). § Backhaul costs can be reduced if self-provided by the operator. For example, KDDI in Japan uses its national Wi. MAX network to backhaul traffic from its network of 100 000 indoor access points. § Alternatively, the premises owner may pay for his own backhaul, while the wireless ISP levies an annual charge for service management. 1 Includes new phone line and installation by engineer. © Analysys Mason Limited 2011 More-detailed analysis of costs is given in the Annex to this report.

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 24 The total cost of providing additional capacity using the indoor Wi-Fi only model is EUR 124 910 Figure 13: Cumulative capital investment (capex and opex) in indoor Wi-Fi only model [Source: Analysys Mason, 2011] § The cumulative spend in the indoor Wi-Fi only model increases gradually over the period 2012– 2018. § A steady rise of relatively small increments has a positive impact upon the MNOs business, as capex and opex spend can be more closely matched to demand. § In order to avoid exceeding the capacity limits on the 3 G site, we have made the reasonable assumption that the operator will direct indoor users onto the indoor Wi-Fi network. It is not clear if the operational support systems can currently support this facility. § Over and above the cost of borrowing and the simple fact that Wi-Fi deployment costs are less than those of the equivalent LTE, indoor Wi-Fi has two further, subtle benefits: - indoor users are typically heavy data users. They will generate less network interference and receive higher data rates when using an indoor Wi-Fi connection than outdoor 3 G or 4 G - the number of indoor users that the outdoor macro network needs to support will fall increasingly over time. This will allow the outdoor network to manage RF power better for the benefit of outdoor users and will improve the outdoor Qo. S. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 25 Outdoor plus indoor Wi-Fi model § In the outdoor plus indoor Wi-Fi model, the fundamental assumption is again that mobile services are provided to the area by a 3 G radio site, but that the radio site is congested and is not able to support any further traffic. The operator’s solution is to build several outdoor Wi-Fi sites. § For the sake of cost comparison, it would have been consistent to have modelled a pure outdoor Wi-Fi scenario. However, the majority of customers will be found indoors and, as outdoor Wi-Fi will not penetrate effectively into buildings, it is unlikely that outdoor Wi-Fi alone will offload sufficient traffic to prevent the outdoor 3 G site from becoming overloaded as traffic increases. We therefore designed the model incorporating deployment of extra indoor sites. § The model was designed such that the outdoor 3 G site is maintained just below capacity saturation, and the number of outdoor and indoor Wi-Fi access points deployed is sufficient to support the surplus traffic. § Outdoor sites are deployed wherever possible in preference to indoor sites. In effect, this is an outdoor and indoor Wi-Fi coverage model, but in which outdoor Wi-Fi deployments are prioritised. § Unlike the indoor case, in which access points are deployed in single units, we decided to model the outdoor Wi-Fi network according to a mesh architecture, which is discussed in more detail in the next slide. © Analysys Mason Limited 2011

26 The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Our outdoor plus indoor Wi-Fi model assumes a mesh architecture § Mesh architecture has a number of advantages when a large number of access points (APs) are located within a common area. The most important point is that the APs collectively can be used to provide their own backhaul to aggregation points, reducing backhaul costs. Figure 14: Outdoor plus indoor Wi-Fi model architecture [Source: Analysys Mason, 2011] Fixed backhaul connection to Internet Main Ethernet switch § The mesh cross-connects the multi-radio access points and only one in six requires dedicated (local) backhaul. This significantly reduces costs. (In the case of single APs (not meshed), each will require backhaul. A detailed discussion of the ratio of gateways to pure APs can be found in the Annex to this report. ) § In our model we have limited the maximum number of Wi-Fi APs in the 3 G site area to ten. Given that the coverage radius of the area of the model is 500 m, and the coverage radius of each outdoor access point is 0. 16 km (160 m), about ten access points must be deployed (equivalent to a density of about 13/km 2). Two access points act as gateways, also providing access for users. © Analysys Mason Limited 2011 Fixed backhaul Ethernet interface Gateway/ Access point Fixed backhaul Access point Fixed connection Access point Wireless connection

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 27 The total deployment cost for an outdoor plus indoor Wi-Fi mesh network is EUR 61 850 and total recurring costs are EUR 22 900 per annum Figure 15: Total deployment costs for an outdoor plus indoor Wi-Fi mesh network [Source: Analysys Mason, 2011] § For ten access points, the total deployment cost is EUR 61 850. § Hardware costs, at EUR 31 700, dominate, representing 51% of total, due to the use of relatively costly multi-radio APs. § Installation costs contribute EUR 15 400 (25% of the total). § Backhaul costs add EUR 7950 (13% of total). 1 § Pre-plan, commission and integration add a modest EUR 2750 (4% of total) because such activities are more straightforward than for a 3 G network. § The total recurring costs are EUR 22 900 per annum. Figure 16: Total recurring costs for outdoor plus indoor Wi-Fi mesh network [Source: Analysys Mason, 2011] § Site rentals contribute EUR 11 000 (48% of total). This value is the least certain, as rental varies widely by location. Agreements with public authorities can reduce this cost significantly. 2 § Although a mesh network topology has been used to minimise backhaul costs, they still comprise EUR 5430 (24% of total). Backhaul costs can be reduced if self-provided by the MNO. § Programme management costs add EUR 1090 (5% of total), while maintenance and power costs contribute 24% of total. 1 Includes: one local backhaul, two gateways to switch and one Internet connection. We have used list prices for backhaul circuits, but contract prices may be lower. 2 In September 2011, Virgin Media held discussions with local councils about developing a Wi-Fi network around London. . More-detailed analysis of costs is given in the Annex to this report. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 28 The total cost of providing additional capacity using first outdoor, then indoor Wi-Fi cells is EUR 260 317 Figure 17: Cumulative capital investment (capex and opex) in outdoor plus indoor Wi-Fi model [Source: Analysys Mason, 2011] § In the first year, 2012, only four external access points will be required to support traffic increases. § In the second year, 2013, the number of access points necessary will rise quickly to ten. § At this point, the outdoor area is saturated with Wi-Fi access points. § Each year thereafter, an increasing number of indoor access points are required to carry the ‘overflow’ traffic from the outdoor macro cell. § The increase in accumulated spend after 2013 is caused by the capex and opex for the indoor access points and opex for the outdoor access points. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks The cost of capacity required to relieve congestion in a 3 G macro-cell is lowest for indoor Wi-Fi only § We have compared the annualised costs of our three deployment models: LTE only; indoor Wi-Fi only; and outdoor plus indoor Wi-Fi. § The total cumulative costs for indoor Wi-Fi and for outdoor plus indoor Wi-Fi are both significantly lower those for 4 G upgrade in the LTE-only model: EUR 124 910, EUR 260 317 and EUR 623 431, respectively. § The indoor costs are the lowest and they increase in a gradual fashion. This is in itself offers a distinct advantage. § It is possible that the cost of outdoor portion of the outdoor plus indoor Wi-Fi deployment model could be reduced by deploying fewer access points, and concentrating them in areas of micro-congestion within the cell. § It is important to remember that, despite the low cost of its deployment, Wi-Fi has a number of operational disadvantages that, while difficult to quantify, will have an impact on the total cost of ownership. These points are discussed in the final section of this report. © Analysys Mason Limited 2011 Figure 18: Comparison of cumulative costs for three deployment models [Source: Analysys Mason, 2011] 29

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 31 In the near future, next-generation Wi-Fi hotspots will be an important strategic element in MNOs’ churn reduction § Wi-Fi hotspots represent a big market, with many big players – notably large hotspot aggregators. Nonetheless, MNOs occupy an important place in the value chain. § Recently, MNOs have been highly active in this space, and many have started to implement their own hotspot deployments and to agree commercial arrangements for hotspot ‘roaming’. This has been motivated by customers’ expectations of being able to access Wi-Fi hotspots at certain locations outside the home. § Our analysis of cost benefits shows that MNOs should consider making their own Wi-Fi deployments, possibly complementing these with partnerships in order to increase their hotspot footprints. It should, however, be remembered that having entered into partnership, the MNO no longer has complete control over managing the user experience and that bad experiences can damage users’ relationships with the operator, leading to churn. § Recent developments support the claim that Wi-Fi is on the verge of becoming an integral network component for MNOs. This is driven by the emergence of next-generation hotspots (NGH), also referred to as Hotspots 2. 0, which signal the end of Wi-Fi’s role as a complementary component of MNOs’ service offerings. They will provide a better user experience and should result in greater customer loyalty and reduced churn. © Analysys Mason Limited 2011

32 The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks The Wi-Fi hotspot value chain encompasses a number of commercial relationships Figure 19: The Wi-Fi hotspot value chain [Source: Analysys Mason, 2011] Global hotspot aggregators National, local and regional hotspot operators Network operators Proprietors of hotspot premises Arrows represent commercial relationships End users § Wi-Fi technology has advanced beyond its initial usage model of providing indoor, short-range wireless Internet access. In the last decade or so, many companies have established large networks of Wi-Fi access points, or hotspots, in public spaces. The value chain shown above describes the many commercial considerations of the Wi-Fi hotspot proposition. § The hotspot provision market has been stimulated by a rapid reduction in the cost of Wi-Fi components brought about by economies of scale driven by an insatiable demand for data services. § As an indication of the growth in this market: - in March 2003, there were only about 3000 commercial hotspots in the USA - current best estimates suggest that there are 225 000– 250 000 hotspots globally - the Wireless Broadband Alliance (WBA) predicts an increase of 350% in global Wi-Fi hotspots by the year 2015. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 33 Commercial entities, such as i. Pass, that aggregate public Wi-Fi hotspots are at the top of the value chain § Characteristics of global hotspot aggregators: - huge footprints. Unlike cellular networks, which are restricted to certain countries and geographical regions, global hotspot aggregators have huge footprints Figure 20: Number of access points accessible to the largest hotspot aggregators [Source: Analysys Mason, 2011] - innumerable commercial agreements. Global hotspot aggregators have commercial partnerships with local and national hotspot operators - relative simplicity. One commercial advantage is that subscribers need only a single account to access hotspots in many regions. These customers are likely to be sophisticated users, familiar with the ‘choreography’ of hotspot network connection - symbiotic relationship with MNOs often have agreements in place with these firms, rather than negotiating many individual ‘roaming’ agreements - but the costs may be high. We have heard that the cost per gigabyte of using the services of some aggregators may be as much as ten times that of self-build. © Analysys Mason Limited 2011 § At the top of the value chain are commercial entities that own, operate and aggregate public Wi-Fi hotspots in a number of countries. As well as operating hotspots, these firms have built up a network of hotspots through commercial partnerships. § Over the last decade, a number of these operators have accumulated a global critical mass, and consequently there are only a handful of serious international players.

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 34 National, local and regional operators of Wi-Fi hotspots § These entities have in place long-standing exclusive commercial agreements with airports, hotels, convention centres, cafés and bars to provide Wi-Fi hotspots at these locations. § A good example is The Cloud is Europe’s largest public Wi-Fi network operator, with a network of 22 000 hotspots in 7000 locations in 12 European countries. It had registered more than 600 000 unique users in Europe by July 2010. § The Cloud also has partnerships with other Wi-Fi providers, such as Telenor, Sprint PCS, AT&T and Tele 2, and with the global aggregator i. Pass. § The Cloud targets three types of client: - businesses, offering wireless broadband to different verticals, such as hospitality, retail, transport, restaurants, hotels and government and local authorities - service providers, building and managing Wi-Fi networks for mobile operators, device manufacturers, network providers, content providers and application providers - end users, offering access to Wi-Fi in specified locations to consumers and business users on pay-monthly or prepaid schemes. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Operators may form partnerships to expand their hotspot footprints, but should complement this with their own deployments § The short range of Wi-Fi networks means that no single MNO can build all of the hotspots necessary to meet the high levels of coverage demanded by customers. It may be wise for MNOs to enter into partnerships with hotspot aggregators for a number of reasons: - presence: partnerships give MNOs extensive hotspot footprints - market entry: partnerships give MNOs immediate access into the market - finance: partnerships reduce capex and opex by outsourcing the service provision. § However, MNOs must address a number of issues, if considering making agreements with hotspot aggregation firms to provision their primary hotspot networks: - quality control: the MNO has little or no control over the Qo. S provided by the aggregator. This may be controlled to a certain extent by NDA agreements, but is not ideal for MNOs committed to managing the user experience - network management: the MNO will have no access to vital hotspot network statistics that it could use to optimise its service and subscriber package offerings. Furthermore, it will not have the power to decide hotspot locations - knowledge deficit: the MNO will lack access to hotspot network statistics that it could collate to better understand its subscribers activities. © Analysys Mason Limited 2011 35

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 36 In the UK, only one MNO has so far announced plans to deploy its own Wi-Fi network § The players in the UK Wi-Fi market are: - Telefónica UK (O 2), which recently announced the launch of a new network – ‘O 2 Wi-Fi’. O 2 aims to have 14 000 hotspots throughout the UK by 2013, and claims that the service will be free to all users. In October 2011, the company announced plans to hold a Vo. IP trial that allows smartphone users to use voice and text services over Wi-Fi networks. O 2 is the only MNO in the UK to have announced plans to deploy its own Wi-Fi network - BT offers its own Wi-Fi service – BT Openzone. The subscription service gives users access to hotspots operated by BT and those owned by partner providers. The company also operates the BT FON service, through which BT Wi-Fi broadband customers share their home broadband with other subscribers. BT operates outdoor hotspots in major cities, but also provides indoor installation and service management services to retail businesses. It expects to have more than 2. 5 million hotspots in operation by the end of 2011 - BSky. B operates about 20 000 hotspots in the UK, achieving this through acquisition of The Cloud, which has struck deals with a number of chain retail outlets. BSky. B will use Wi-Fi to offer high-quality mobile streaming video services to its subscribers § Virgin Media has no Wi-Fi footprint, but in September 2011 met with local councils to consider deployment of a Wi-Fi network around London. Agreements with public bodies can minimise potentially costly site rental charges for outdoor access points. § The rapid and widespread proliferation of hotspots in locations in the UK with high footfall is expected, and this will further stimulate the use of mobile broadband. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 37 Elsewhere, a number of large-scale Wi-Fi roll-outs have recently been announced § Towerstream in the USA is currently building an outdoor Wi-Fi network of 1000 access points across seven square miles of Manhattan. 1 Backhaul will be self provided, using the company’s existing rooftop point-to-multipoint wireless network. It will sell capacity to carriers that need to offload traffic from congested cellular networks. § In September 2011: - KDDI, Japan’s second-largest MNO, announced plans to roll out a national network of 100 000 indoor Wi-Fi hotspots by March 2012. The network will be used largely to relieve pressure on its cellular network, but will also offer subscribers a fast Wi-Fi service. Traffic will be backhauled over KDDI’s existing national Wi. MAX network - France Telecom announced that Wi-Fi was its preferred option for data offloading, and plans to use its 7 million DSL boxes with Wi-Fi and the 40 000 public Wi-Fi hotspots in service in France - Time Warner Cable in the USA announced USD 15 million plans to roll out a Wi-Fi network across 40 square miles of Southern California. The main driver is to offer free Wi-Fi access to its existing subscribers; non-subscribers will be able to access the service for a fee - BSky. B’s Wi-Fi subsidiary in the UK , The Cloud, announced multi-million pound plans both to upgrade and significantly to expand its network of UK Wi-Fi hotspots in order to provide offloading capacity to MNOs. § MNOs that have not yet entered the Wi-Fi hotspot space are likely to make large, rather than incremental, deployments. 1 The access points are based upon the latest 802. 11 n wireless standard, which builds on previous 802. 11 standards by adding multiple-input multiple-output (MIMO) and 40 MHz channels to the PHY (physical layer), and frame aggregation to the MAC layer. 802. 11 n offers superior range and data throughput and is more suited to the delivery of IP streaming video than its predecessor, 802. 11 g. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 38 Recent developments support the claim that Wi-Fi is on the verge of becoming an integral network component for MNOs § AT&T and the WISPr protocol: Apple i. Phones and Windows Phone 7 phones from AT&T have WISPr protocol support. This allows phones to switch automatically from cellular data to AT&T’s Wi-Fi hotspots, where they are available. § Deutsche Telekom and i. Pass: Deutsche Telekom views Wi-Fi as ‘a re-emerging technology’. In partnership with i. Pass, it has launched ‘Wi-Fi Mobilize’, a solution that incorporates a software client on user terminals to act as a smart agent for connection management and network selection. § KDDI and Ruckus Wireless: KDDI is to roll out 10 000 indoor Wi-Fi hotspots in public venues throughout Japan using Wi-Fi infrastructure based upon the 802. 11 u standard supplied by Ruckus Wireless. § Republic Wireless: This US company offers a hybrid phone that allows calls to be made using any available Wi-Fi hotspot, falling back to the mobile network when no Wi-Fi is in range. The company estimates that most people are near a Wi-Fi network for 60% of the time. § i. Pass OMX: i. Pass has activated Open Mobile Exchange (OMX) which claims seamless ‘zero-click’ authentication and ‘roaming’. OMX is a software platform developed by i. Pass that allows MNOs to connect mobile customers to Wi-Fi networks in a way that is transparent to the customer. § Shaw Communications and Cisco: In October 2011, Shaw Communications in Canada and Cisco announced that they would conduct a technical trial of Hotspot 2. 0. Shaw owns Advanced Wireless Service (AWS) spectrum, but recently shelved plans to become a traditional MNO in favour of deploying a Wi-Fi hotspot network. § Various: Various MNOs have launched Wi-Fi offload apps for smartphones and tablets. These increase awareness of hotspot availability and allow subscribers easily to locate local hotspots. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Operators worldwide are busy with hotspot deployments [1] Figure 21 a: Examples of recent, large-scale deployments of Wi-Fi by MNOs [1] [Source: Analysys Mason, 2011] Operator Country Number of access points Comments 3 BB Thailand 15 000 None AT&T USA 24 000 None Beeline Russia 14 000 (Moscow) None China Mobile, China Unicom and China Telecom China 90 000 Plans for another 390 000 hotspots by 2016 du UAE 92 None KDDI Japan 10 000 Plans for a total of 100 000 hotspots in conjunction with Ruckus Wireless KT South Korea 87 000 None LG Uplus South Korea 40 000 None Oi Brazil 40 500 40 000 hotspots acquired through acquisition of Vex © Analysys Mason Limited 2011 39

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 40 Operators worldwide are busy with hotspot deployments [2] Figure 21 b: Examples of recent, large-scale deployments of Wi-Fi by MNOs [2] [Source: Analysys Mason, 2011] Operator Country Number of access points Comments Orange France 30 000 None PCCW Hong Kong 9000 None SFR France 30 000 Claims access to 3 000 hotspots worldwide Shaw Communications Canada 0 Has owned AWS spectrum since 2008, which it will retain as an asset. Has now shelved plans to deploy traditional network in favour of Wi-Fi, which should save it USD 1 billion SK Telecom South Korea 51 000 None Swisscom Switzerland 1350 Claims access to 65 000 hotspots worldwide. Also operates 2000 hotel hotspots across Europe (Eurospot) Telmex Mexico Over 4000 None Telstra Australia 49 Plans for a wider scale deployment VIVACOM Bulgaria 2000 None © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 41 Customers have come to expect that certain commercial premises will have Wi-Fi access points and this helps to stimulate the market § In many developed and emerging markets, cafés, bars and restaurants now offer Wi-Fi Internet access and many retail stores also provide direct unrestricted access to the Internet. Initially, these were services for which the patron would have to pay. Now, however, there is an established trend for retailers and restaurants to migrate from a paid-for to a free model for the provision of wireless access. The declining cost of mobile broadband has made it impractical to charge for Wi-Fi. § The main objective is to attract and retain customers who, if they are able to access the Internet on the premises, will consume more onsite. Also, by asking customers who use the free Wi-Fi access to register their details, it is possible to collect important information about Internet usage habits and to use direct marketing. § In order to provide Wi-Fi connectivity in their premises, companies can enter commercial agreements with either traditional network operators or hotspot aggregators. Mc. Donald’s has taken both approaches: - In the UK, Mc. Donald’s began to offer free Wi-Fi, supplied by The Cloud, in its 1200 restaurants in October 2007. The Cloud also provides corporate connectivity services for the restaurants’ back-office functions, such as till reporting - In the USA, Mc. Donald’s has offered free Wi-Fi since January 2010 in 10 400 of its 14 000 restaurants. The network was built, and since 2003 has been managed, by AT&T Wireless and was initially a paid service. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 42 The emergence of NGH signals the end of Wi-Fi’s role as a complementary component of MNOs’ service offering § Wi-Fi hotspots are an important part of operators’ offload solutions, and will continue to be so. The Wireless Broadband Alliance (WBA) predicts a 350% increase in global Wi-Fi hotspots by the year 2015. § In the near future, there are plans to integrate Wi-Fi hotspots seamlessly into MNOs’ macro networks. Heterogeneous networks – those in which different cell sizes and associated technologies (macro, micro, pico and small cell) are combined, allowing handover from one cell to another – will become vital for MNOs because of the cost of meeting consumers’ demands for capacity. § IEEE 802. 11 u is an important innovation in standards. This protocol allows a Wi-Fi base station to appear to the 3 GPP network as a typical 3 GPP base station. For all operational purposes, the Wi-Fi access point appears to the network as a typical 3 GPP small cell (femtocell) device. § Wi-Fi Alliance’s Hotspot 2. 0. Certification software that would automate connection to public hotspots is planned. This will essentially be a SIM-authentication solution like that used in typical cellular communication. It will build upon and extend the IEEE 802. 11 u standard. The solution will be trialled in the third quarter of 2012, and widespread deployment is expected in 2013. § WBA’s Next Generation Hotspots. The WBA is an industry group of vendors and operators, the aim of which is to bring a 3 G-like experience to Wi-Fi, and to make it an integral part of service providers’ radio access networks, on a par with 2 G, 3 G and LTE. The WBA is particularly focused on the issue of interoperability, intending that future Wi-Fi hotspots should allow roaming. § In June 2011 – the WFA and WBA announced that they would harmonise their efforts towards developing carrier-class Wi-Fi, in practice making the terms Hotspot 2. 0 and NGH synonymous. § In the next two years, MNOs with hotspot networks may be expected to begin to upgrade to offer Hotspot 2. 0 capabilities. Furthermore, advances in seamless authentication are being mirrored by technical advances, including Amped Wireless’s Wi-Fi repeater, which extends the range of Wi-Fi transmission to about 1. 5 miles, and Wi. Gig, which allows transmission rates of 7 Gbps in the 60 GHz spectrum band. Innovations like these may also be incorporated into MNOs’ future Wi-Fi offload strategies. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 44 Wi-Fi is cheaper than LTE and operates in free spectrum, so is there a catch? § Moving towards a carrier class of service presents the operator that wishes to deploy Wi-Fi with a number of operational challenges. § As the previous analysis shows, Wi-Fi is a cost-effective delivery mechanism, but it is unclear how many macro sites will benefit from Wi-Fi as a capacity solution. It is evident that only congested cells will benefit. Traffic forecasts tend to average mobile traffic over the network, to show all sites carrying an equal load. In practice, some sites will be heavily overloaded and others will have a surplus of capacity. It will be crucial to the successful deployment of Wi-Fi as a capacity solution to identify the overloaded sites in order to provide extra Wi-Fi capacity. § As with all small-cell solutions Wi-Fi is burdened by the time and cost of site acquisition and hampered by local regulation. Generally, operators do not have the operational capability to roll-out tens or hundreds of thousands of Wi-Fi hotspots. There is a need for service aggregators like i. Pass, or operational partnerships, like that of Telefónica and BT in the UK. § In order to deliver a carrier class of service for mobile operators, Wi-Fi has to overcome three hurdles: - the lack of standardised end-to-end policy control - the lack of standardised SIM authentication - the lack of a standardised means of encrypting transmission. § As it stands, Wi-Fi is not a technology that can provide carrier class services. A number of mobile operators are already using Wi-Fi at the moment where SIM authentication is available, establishing its own non-interoperable authentication solution with its mobile device vendor(s). © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 45 Connecting to a Wi-Fi hotspot should be as simple as connecting to a mobile network and problems may harm the user’s bond with the operator § A core churn-reduction strategy of many operators is to build a consumer-orientated experience. Presently, Wi-Fi hotspots do not entirely live up to this goal, as connecting to them is not easy, seamless or necessarily secure. Problems include: - cumbersome credential entry: customers must log in using passwords and user names: subscriber credentials. This can be an unfamiliar process and burdensome. - log-in to a network via a public Wi-Fi hotspot often requires a sequence of actions to be carried out: after selecting and connecting to a hotspot, users may need to launch a browser in order to enter their credentials. If this is not done, other networkreliant applications will not run, even though the device displays a working network connection icon - requirement for credit card: in the case of pay-as-you-go access, many users are reluctant to give details of their credit cards, or may not own a card - roaming availability: roaming onto other MNOs’ hotspots is rarely supported - SSID spoofing: phony access points, which can be used to steal credit card details or passwords, or to spread malware, are a serious security issue - malfunctions: hotspots may not be designed with sufficient capacity or may quickly become popular and overburdened, resulting in poor or unusable connections. § MNOs must be aware that inconvenience or obstacles in access may negatively affect the user experience and potentially lead to churn. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 46 The Wi-Fi industry must standardise in order to deliver carrier-grade products and services § In order to drive global adoption of Wi-Fi by mobile operators, the areas of authentication, encryption and policy control must be standardised - the WBA, which comprises a group of mobile operators, vendors and other stakeholders, was created to support the adoption of Wi-Fi as an ‘integral component of ubiquitous and seamless wireless broadband services’ - by the end of 2011, the WBA expects to have undertaken user trials of standard mobile devices authenticated and encrypted over Wi-Fi using standard protocols. It expects to see commercial adoption of these standards by mobile operators by the end of 2012. § Some of the issues faced by Wi-Fi are overcome by femtocell technology. However, femtocells have not benefitted from significant economies of scale and are not yet priced competitively. Indeed, Telstra recently rejected the technology, claiming it had been made ‘redundant’ by the widespread adoption of home Wi-Fi. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 47 Backhaul is still a challenge for all small-cell solutions, not only for Wi-Fi § Operators are designing tomorrow’s backhaul based upon yesterday’s architecture. § Current backhaul is not sufficiently finely grained to support the number of offload nodes, which may be as many as 1000 per square kilometre. The points of aggregation are unlikely to be placed in the locations with the greatest probability of supporting Wi-Fi. § It is likely that there will be 50– 100 m gaps between the Wi-Fi access units and the backhaul points of presence. Technologies will be needed to bridge this gap. Novel solutions, like broad-beamwidth (point-to-multipoint) microwave, LTE backhaul and Wi-Fi mesh architecture may be used to bridge the few hundred metres to the first fixed point of presence. § Indoor and outdoor Wi-Fi backhaul have different characteristics. As shown in the analysis, Wi-Fi is best deployed as an indoor solution. For indoor deployments, the provision of backhaul should not present a problem, given the availability of fixed lines (copper and fibre). The challenge for the operator in this case is to access and acquire the site. This may mean leasing access to the site from the site owner, or establishing a commercial relationship with the site owner, for example a chain of cafés. § It is likely, therefore, that MNOs will have to pursue two different business models, driven by the respective backhaul requirements: business-to-consumer (B 2 C) for outdoor Wi-Fi and business-to-business (B 2 B) for indoor Wi-Fi. § The benefits of Wi-Fi as an alterative data delivery architecture for MNOs is very sensitive to the cost of backhaul. This backhaul may include a wireless component to bridge the gap. If governments perceive auctioning this spectrum as a means of increasing revenue, this will probably kill the business case for Wi-Fi offload, if it relies upon a wireless backhaul component. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 48 MNOs will need the services of trusted partners to backhaul Wi-Fi § Much of the wireless traffic to and from web servers does not need to traverse the operators backhaul and core network. Operators are seeking to reduce this traffic by rerouting it to its final destination, avoiding the mobile operator’s network. This is known as selective IP traffic offload. 1 § Typically this will take place in the backhaul part of the network that is controlled by the fixed operator. Deep packet inspection is required to select and reroute the appropriate traffic. This requires that the mobile operator establishes a partnership with the fixed operator. § However, the fixed operator could use knowledge gained in this process to push advertising to the mobile customer, establish a commercial relationship with the customer and even perhaps persuade the customer to download an application that allows the fixed operator to retain ‘control’ of the customer, once the customer leaves the Wi-Fi network and roams back onto the pure mobile network. 1 It is arguable whether SIPTO traffic is mobile or fixed. Indeed, this is a particular thorny issue with which the industry has yet to grapple: finding new industry definitions for mobile and fixed traffic which allows for clear and concise recording and reporting of traffic data. In order to move forward with this problem and not get bogged down by taxonomy, we have defined this as mobile traffic; but the plausible counterarguments that this is fixed traffic are noted. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 50 Indoor Wi-Fi only model assumptions: breakdown of total deployment and recurring costs Figure 1 A: Total deployment costs for an indoor access point [Source: Analysys Mason, 2011] Installation Hardware © Analysys Mason Limited 2011 Figure 2 A: Total recurring costs for an indoor access point [Source: Analysys Mason, 2011]

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 51 Outdoor Wi-Fi unit assumptions: breakdown of total deployment and recurring costs Figure 3 A: Total deployments costs for outdoor Wi-Fi network [Source: Analysys Mason, 2011] Figure 4 A: Total recurring costs for outdoor Wi-Fi network [Source: Analysys Mason, 2011] Backhaul Installation Backhaul Hardware © Analysys Mason Limited 2011

52 The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Mesh networks in more detail [1] § Outdoor networks do not provide reliable indoor coverage, and so separate indoor networks are often deployed. Figure 5 A: Outdoor Wi-Fi model architecture [Source: Analysys Mason, 2011] Fixed backhaul connection to Internet § Figure 5 A shows a common architecture for a Wi-Fi mesh topology. Main Ethernet switch § Outdoor Wi-Fi networks are increasingly configured in a mesh topology Fixed backhaul - mesh topologies minimise the need for costly fixed backhaul connections Ethernet interface Gateway/ Access point - mesh protocols determine the best route through the network - each mesh node forwards packets to other nodes as required - if a node is connected to at least one other node, it has full connectivity to all nodes in the mesh. Ethernet interface Gateway/ Access point Fixed backhaul Access point § Indoor networks usually connect from the premises to the Internet via fixed-line backhaul. Fixed backhaul © Analysys Mason Limited 2011 Indoor access point Fixed backhaul Various types of indoor space (cafés, hotels etc. ) containing mobile devices

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 53 Mesh networks in more detail [2] Figure 6 A: Illustrative average access point capacity versus number of access points [Source: Analysys Mason, 2011] § Mesh networks use access points that can be classified as single, dual or multi-radio. § In a single-radio mesh (the cheapest option), the same radio is used for access to subscribers and for backhaul between access points. As more nodes are added, a greater percentage of the cell capacity is dedicated to backhaul capacity (forwarding of packets) and a smaller percentage to access capacity § In a dual-radio mesh, each access point has separate radios for access and backhaul, with each operating at a different frequency. Access capacity is not reduced by forwarding of packets. However, the backhaul network is still contended by access points. 1 § In a multi-radio mesh, there are separate radios for access and backhaul, and the backhaul links operate in different channels (the most costly option). 1 © Analysys Mason Limited 2011 In September 2011, BSky. B’s Wi-Fi service provider, The Cloud. announced a technologyrefresh plan to replace its legacy network of 802. 11 g access points with a dual-radio 802. 11 n version. This will enhance wireless streaming video services for Sky subscribers and also provide mobile offloading services for MNOs.

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 55 About the authors Terry Norman (Principal Analyst) is the lead analyst for Analysys Mason’s Wireless Networks and Spectrum research programmes. His primary areas of specialisation include wireless access technologies, wireless traffic forecasting and spectrum-related matters. Terry has more than 30 years' experience in the radio communications industry. He has worked in radio planning, access network design, operations and for ten years was a Consultant with Analysys Mason. As well as writing numerous reports on the subject of wireless access technologies, Terry has been quoted in the broadsheet press and regularly contributes to media analysis. He holds an Engineering Doctorate in radio propagation, is a member of the Institute of Physics, a member of the Institute of Technology and a Chartered Physicist. Richard Linton (Consultant) has more than twenty years’ experience in the telecoms industry, and has worked as a Consultant with Analysys Mason for over eight years. He specialises in technical and commercial work for operators in both the fixed and mobile areas. In the mobile sector, he has worked overseas on many projects for a diverse range of clients, from large established operators to small start-up companies. As well as undertaking a number of research assignments, he has contributed to many strategic and operational consulting projects. Richard main skills include cost modelling, market analysis and technical due diligence. He holds a first class honours degree in Physics and a Ph. D in Engineering. Published by Analysys Mason Limited • Bush House • North West Wing • Aldwych • London • WC 2 B 4 PJ • UK Tel: +44 (0)845 600 5244 • Fax: +44 (0)845 528 0760 • Email: research@analysysmason. com • www. analysysmason. com/research • Registered in England No. 5177472 © Analysys Mason Limited 2011. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means – electronic, mechanical, photocopying, recording or otherwise – without the prior written permission of the publisher. Figures and projections contained in this report are based on publicly available information only and are produced by the Research Division of Analysys Mason Limited independently of any clientspecific work within Analysys Mason Limited. The opinions expressed are those of the stated authors only. Analysys Mason Limited recognises that many terms appearing in this report are proprietary; all such trademarks are acknowledged and every effort has been made to indicate them by the normal UK publishing practice of capitalisation. However, the presence of a term, in whatever form, does not affect its legal status as a trademark. Analysys Mason Limited maintains that all reasonable care and skill have been used in the compilation of this publication. However, Analysys Mason Limited shall not be under any liability for loss or damage (including consequential loss) whatsoever or howsoever arising as a result of the use of this publication by the customer, his servants, agents or any third party. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 56 About Analysys Mason The only constant is change. What worked yesterday won’t necessarily work today. That’s why we look beyond the obvious, seeing things from a client’s perspective so that a truly effective solution is delivered every time. A key part of this is our international perspective. Business never sleeps, and with offices spanning six time zones, neither does Analysys Mason. Telecoms, media and technology (TMT) are our world; we live and breathe TMT. This total immersion in our subject underpins and informs everything we do, from the strength and reliability of our market analysis, to improving business performance for clients in more than 100 countries around the world. At the core of our approach is a simple, but enormously powerful idea: applied intelligence. By harnessing our collective brainpower we can solve real-world problems and deliver tangible benefits for our customers. As a Japanese proverb says, ‘all of us are smarter than any of us’. We’re passionate about what we do, with the focus and determination to take on and solve the toughest problems to help our clients. We’ll rise to the challenge and enjoy it. In fact when it comes to problem solving, there’s a real sense of ‘the tougher the better’. It’s this unique combination of our applied intelligence, effective problem solving and the ability to look closer and see further that makes Analysys Mason special. © Analysys Mason Limited 2011

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks Research from Analysys Mason Our subscription research programmes address key industry dynamics in order to help clients interpret the changing market. The programmes focus on five key practice areas: We analyse, track and forecast consumer and enterprise services, as well as the software, infrastructure and technology that underpins the delivery of those services. Subscribing to our research programmes gives you regular and timely intelligence. It also provides direct access to our team of analysts – that is, the opportunity to engage one-to-one with our subject experts for insight, opinion and practical advice relating to your most-critical business decisions. To find out more, please visit www. analysysmason. com/research. © Analysys Mason Limited 2011 57

The case for Wi-Fi offload: the costs and benefits of Wi-Fi as a capacity overlay in mobile networks 58 Consulting from Analysys Mason For more than 25 years, our consultants have been bringing the benefits of applied intelligence to enable clients around the world to make the most of their opportunities. Unlike some consultancies, our focus is exclusively on TMT. We advise clients on regulatory matters, support multi-billion dollar investments, advise on network performance and recommend commercial partnering options and new business strategies. Such projects result in a depth of knowledge and a range of expertise that sets us apart. We blend our range of skills each day, every day, to solve our clients’ most-complex challenges. Our skill set is broad. It has to be. Our clients in the TMT sectors operate in dynamic markets where change is constant. We help shape their understanding of the future so they can thrive in these demanding conditions. To do that we have developed rigorous methodologies that deliver real-world results for clients around the world. To find out more, please visit www. analysysmason. com/consulting. © Analysys Mason Limited 2011