The explosive growth in video consumption on mobile devices is a huge elephant in the room affecting the course of 4G/LTE deployments and the auctions of spectrum that precede them.
Europe is lagging behind the U.S. and Far East over 4G deployment, with the auction season currently in full swing and generating some interesting results. Roll back just over a decade to the auctions of 3G spectrum, and governments were hell bent on making as much money as possible with competitive aspects almost taking second place. The UK raised £22.5 billion ($34 billion at current exchange rate) from that sale in 2000 at a time when prices in all places were grotesquely inflated by the dot-com boom coupled with unrealistic expectations of how much consumers would be prepared to pay for mobile data. Video was not then in the equation, nor was mobile telephony considered to be an essential service that should not be too expensive.
4G spectrum prices are much lower reflecting more realistic current expectations, with the recently completed UK auction fetching £2.34 billion, little over a tenth of the 3G one. This is all well known, but the interesting aspect of some European auctions, raising eyebrows elsewhere in the world, is that they have been conducted more with a view to stimulating competition and keeping down the price of the services eventually deployed, than on raising money. This could all have significant implications for future mobile video and TV services.
The Czech Republic’s case really made this point, since the CTO (Czech Telecommunication Office) actually halted the country’s multi-band auction for spectrum at 800MHz, 1800MHz and 2.6GHz frequencies early this month because the prices had got too high. This seemingly perverse decision where a government acts to deny itself revenue echoed the UK situation, where the £2.34 billion raised was deemed disappointing compared with the government’s estimate of £3.5 billion. Ofcom, the UK regulator, insisted though that the auction had not been constructed to maximize price but to stimulate competition, for example by ensuring that one license went to a wholesale provider that would in turn bring in other brands to offer retail services, perhaps specializing in video or TV.
This Czech case begged the question of whether prices were getting too high there. According to Mark Colville, Senior Manager at Analysys Mason, auction prices had reached CZK20 billion ($1.025 billion), equating to $97.5 per head of population, or given that a total of around 300MHz of spectrum was being offered, $0.325 per MHz/head, which is a common yardstick for price comparison between countries.
As Colville noted this is about 30-percent higher than the UK, where the auction raised about $0.22 per MHz/head, but under half the $0.82 raised in the Netherlands auction late in 2012, where the focus appeared still to be more on maximizing revenues. This led Colville to speculate the CTO might be wrong in assuming that high spectrum costs automatically translate into excessive prices for consumers, particularly 4G internet services, because this assumes lack of competition as well.
As the 3G situation showed operators struggled to recoup spectrum costs in a competitive market, and the bigger issue, as highlighted by the European Union’s Vice President for Digital Agenda Neelie Kroes at the recent Mobile World Congress in Barcelona, was that high spectrum prices would delay the time to deployment because it would take longer for operators to stump up the cash for the required infrastructure upgrade.
This is where WiFi comes in. Although it has hardly been mentioned as a factor in spectrum auctions by analysts, the growing availability of WiFi as an option for offloading mobile data from the cellular networks onto the broadband infrastructure, which generally offers much greater economies of scale and potential capacity than the backhaul networks serving cell towers around the world, is having a growing impact on mobile data plans and pricing. Operators are relying increasingly on WiFi to offload data onto broadband, but at the time some are concerned that this will reduce revenue from their 3G or 4G networks.
The issue is not about comparing 4G with WiFi, since their relative merits in terms of data rate and also device battery consumption depend almost entirely on the number of users per cell, the amount of traffic per user, and the size of the cell. Data rates and battery efficiency fall off with distance from the nearest cell tower or WiFi access point, irrespective of traffic levels.
In terms of spectral efficiency 4G is at least as efficient as WiFi and where it has been deployed has led to at least a temporary dip in WiFi traffic as users find they can enjoy watching video on the move rather than having to wait until they are within a WiFi hot spot. In South Korea, where 4G services are already in full throttle, a survey by Mobidia, a vendor of software for enhancing mobile network performance, showed that 4G users were consuming 2.2 GB of data per month, compared with 1.3GB for 3G users, a big jump.
The problem with LTE is almost that it works too well, and drives down cell sizes to meet growing radio capacity demands, which transfers the pressure onto the backhaul network. This led Suk-Chae Lee, head of Korea Telecom, to admit at the Mobile World Congress in an interview with Reuters, that LTE was making his users“data crazy”, although he was referring largely to video.
He highlighted the problem as being that consumers were not prepared to pay for the necessary fixed line infrastructure upgrade, leading him to conclude rather helplessly, "LTE is very beneficial to the people but still the big question remains: can we go on? It is a blessing to customers but it is a curse on the operators.”
This brings us back to the spectrum auctions and WiFi. The real question is the backhaul, given that to an extend the radio specxtrum issues will be resolved by making cells ever smaller, with further spectral efficiency improvements likely to be relatively modest. In this context we can see that the idea of a 5G beyond 4G is really a red herring when the issue is how most cost effectively to deliver the bandwidth to the cells.
The answer lies with WiFi served by the existing broadband infratsructure, which is being built out anyway, especially as in many countries WiFi hotspots already provide coverage almost matching that of the cellular network. There is work to be done still on handoff, roaming, and authentication, but various initiatives, notably the Wireless Broadband Alliance’s Next Generation Hotspot (NGH) program based on the WiFi Alliance’s Hot Spot 2.0 standard, are moving towards carrier WiFi where hotspots become more cell like and can be integrated with mobile operators’ networks. There is still the question of who pays for the bandwidth, but whatever happens it will be a lot cheaper to hop onto the fixed broadband infrastructure than use cellular backhaul networks. It is true that WiFi is not a free ride and its costs will converge towards those of 4G as the price of carrier WiFi extensions is factored in, but the curves will not meet in the foreseeable future.
This raises the question of whether some operators will delay 4G deployments while they wait and see how carrier WiFi plays out. In some cases at least even the prices paid for 4G spectrum may come to be seen as too high, just as the Czech government feared was happening there.