‘Build and they will come’ was the great fallacy of the early 3G build-outs a decade ago. Consumers did not understand the benefits and did not like the devices; many networks remained underused for years, especially in Europe; and some operators like 3 made more profit from 3G’s improved voice efficiency than from data for many years. The exact reverse has happened in early LTE, where operators have sometimes underestimated consumers’ greed for additional data speed, capacity and quality of service. Verizon CFO Fran Shammo said this week: ‘We did not anticipate that amount of growth in the network’, and both the top two US carriers have been discussing plans to accelerate expansion of their 4G roll-outs next year.
Of course, this is not a global trend. Many cellcos remain cautious about scaling up too quickly, especially those with plenty of rope left in their 3G networks (not the case for the CDMA players which have tended to lead the massive LTE deployments). But in contrast to early 3G, the applications for souped-up mobile data networks are well understood; attractive devices were launched far more quickly than expected; and rather than charging high premiums for 4G, many carriers have used it to lower cost of delivery and to encourage new consumer behaviors, such as accepting tiered data pricing or multi-device bucket plans. These appear to be good value for the user while helping the operator match revenue to usage more accurately.
All these factors, as well as its considerable headstart, have brought Verizon to the point where its LTE networks are already congested in certain areas of high usage, notably the downtown districts of New York, San Francisco and Chicago. ‘There are certain pockets where we’re absolutely going to experience that downtick from the LTE network down to 3G because of capacity constraints,’ Shammo admitted, though he added: ‘By the end of this year you are going to see all those issues dissipate.’
The signs were there early ‘Shammo said on an investor call that the operator had been forced to back away from its previous unlimited data plans more rapidly than expected, because of the scale of demand.
‘What we saw happen when you move from a 3G to a 4G network, the consumption of data increased tremendously,’ he said. ‘And we knew that if a small percent did that, what happens when you get a large percent moving? You have to be able to monetize your capital investment somehow.’
Of course, one issue is that the initial roll-out was geared to broad coverage and took place in the 700MHz band, which is cost-effective for covering sparsely populated areas because of its long range, and is good for indoor penetration (most high end data usage takes place inside). However, the low frequency spectrum quickly runs out of capacity, so Verizon and AT&T will both be accelerating the addition of higher band airwaves into their networks, building up capacity in urban areas and moving towards carrier aggregation. This is a key feature of the LTE-Advanced menu of technologies, which allows two or more bands to be treated as a single resource. Verizon and AT&T both have AWS spectrum, and will start to refarm PCS spectrum as usage of 2G/3G services declines. AT&T has also been experimenting with some more esoteric bands such as WCS (2.3GHz), which it acquired from NextWave, and the unpaired 700MHz frequencies previously owned by Qualcomm.
Shammo said Verizon will boost its capex budget by $500m this year to bolster capacity and speed in key cities. The 2014 focus will be firmly on capacity as the operator has all but completed its coverage, matching its 3G footprint. ‘We’re done with coverage,’ he said. ‘We’ll be in the capacity game.’
AT&T is in the vanguard of using small cells ‘ Wi-Fi, cellular and multimode ‘ to increase capacity in a low cost and targeted way. It has now deployed 3G small cells, indoors and outdoors, across 18 US states to extend HSPA+ coverage and capacity and ease strain on the main network. Next, it will start to introduce ‘multistandard metrocells’ (MSMs) during 2014. These will combine 3G, LTE and/or Wi-Fi and support aggregation of the 700MHz and AWS bands.
It aims to deploy at least 40,000 public access small cells in 2013-2015 and the MSM will eventually become its standard platform. AT&T’s
To maximize the capacity of small cells and avoid interference in dense deployments, optimal positioning is important and automated self-organizing network (SON) tools only address part of the issue. Engineers like those at AT&T Labs are working on RF propagation models to identify the best positions for tiny base stations and some operators ‘ or SCaaS (small cell as a service) providers ‘ believe competitive edge will lie in having the best tools.
AT&T’s inhouse development has turned into a proprietary took called HARP (HetNet Analysis and Resource Planning), which analyzes a host of data points to understand how RF waves behave in a small cell environment. Based upon a location’s traffic demands, along with building plans or topography maps, HARP predicts coverage for both indoor and outdoor small cells and recommends the optimal number and placement.
“It identifies where small cells should be placed and recommends a backhaul solution, all with the goal of delivering an optimal customer experience and maximizing capacity and coverage,” John Donovan, senior EVP of AT&T’s technology and network operations, wrote in a blog entry.
The data points which are fed into the model include “high resolution geolocated demand, macro network RF and performance metrics, high resolution clutter, 3D structure polygons, as well as building location level information.’
According to Sarat Puthenpura – AT&T Labs executive director of service quality management research, and one of the five-strong team which created HARP – the fundamentals of RF propagation models are similar whether the cells are indoor or outdoor, enabling creation of the same planning models for both.
The carrier has also started to refarm 1.9GHz PCS spectrum in certain high traffic areas like central New York City. LTE services in PCS spectrum will launch commercially in Baltimore, Dallas, NYC, Philadelphia, San Francisco and Washington DC by the end of this year and AT&T is preparing the way for a shutdown of 2G, converting users still with GSM/GPRS-only handsets to higher level services. The PCS 4G network is only 2x5MHz so will support 50% the speed of the 700MHz system, which has a total of 20MHz of bandwidth. But with aggregation, this will be a useful extra source of capacity before the major roll-out in AWS stars.