Your browser is not supported. Please update it.

11 March 2008

Shared RANs, femtocells and OFDM soften but don’t solve mobile data problem

As the US cellcos race towards flat rate data as well as voice, there has been rejoicing among the vendors of WiMAX and LTE kit, amid the assumption that operators will need higher capacity, more spectrally efficient new networks in order to sustain the likely increase in traffic and still have any hope of making margin. However, some carriers are finding it hard to balance the additional capex/reduced revenue per Mbps equation, even while lured by the promise of lower churn and greater ARPU. T-Mobile International, often the most astute of the majors in understanding the implications of the open internet model on cellcos, sounded some timely notes of caution during the recent CeBIT trade show in Germany, calling on the industry to reuse its existing networks more efficiently, and to prepare for a day when networks and services will have to be divorced from one another.

Speaking at a meeting of the influential operator-driven Next Generation Mobile Networks (NGMN) group, held at CeBIT, T-Mobile’s CEO Hamid Akhavan pointed out that flat rate tariffs on data meant that network usage would grow far more quickly than operator revenue, putting a huge strain on profitability and the ability to invest in the capex needed to support high capacity networks. Costs would have to go down dramatically, and while spectrally efficient modern networks like WiMAX would help, in the short term there would need to be a stronger focus on using existing assets more effectively, and in the long term, the natural consequence would be to have one or more utilities that owned and ran the networks, leasing capacity to separate service providers ‘ as in other utility sectors like gas and electricity.

“We will see a complete decoupling of traffic and revenues,” predicted Akhavan. “It is only a matter of time before we lose all profitability on mobile data. In the past, user experience has been driven by the average traffic on a cell site. Today it has to do with how much peak traffic you can carry in a cell.” Spectral efficiency, simpler and flatter networks, and reuse of infrastructure would all play a role while operators still held on to their dual role. “The world doesn’t need any more cell sites,’ said the CEO.

This point was taken up by Hartmut Kremling, CTO at Vodafone Germany, in the same debate. He pointed to the difficulty, in terms of cost and environmental regulation, of acquiring cell sites, especially in Europe, and said: “We need a technology based on existing density. Backhaul is already very challenging ‘ increasing the number of sites will exacerbate this dramatically.” The panellists called for plans that would involve LTE reusing current sites for at least 80% of the network, as well as the adoption of more flexible approaches to RAN design, that could reduce dependence on large base stations and tall masts.

These include femtocells, or tiny indoor base stations ‘ T-Mobile has announced an investment in Ubiquisys, a UK start-up that has been a pioneer in this field; and Vodafone, an early enthusiast and triallist, said at its Technology Update in London last week that it could offer commercial femtocell services by the end of the year (in line with O2). The company says femtocells could save it up to 20% in capex terms in extending a 3G network, compared to a macro approach, and early indications are that these economics will be more impressive when applied to a brand new future network such as WiMAX or LTE. Akhavan said at CeBIT that T-Mobile would start LTE roll-out in 2010 and would use femtocells for part of the RAN right from the start, even while accepting that issues connected to interference, power consumption and device cost needed to be addressed conclusively before then.

Vodafone CTO Steve Pusey, speaking at the Update, said: “We’ve approached femtocells from a cost savings perspective. It gives us the potential to improve coverage and reduce backhaul costs. It may or may not have commercial opportunities beyond that.”

Indicating the scale of the issue Akhavan was highlighting, estimates are that mobile traffic will increase a hundredfold worldwide by 2015, and at that stage, over 80% of internet traffic will be handled on mobile networks, with mobile data usage overtaking fixed around 2012. While the content and applications providers ‘ and even more, the mobile device makers ‘ may be rubbing their hands at such forecasts, for the network operators they are definitely a double-edged sword. Network innovations with new economics, such as the data spectral efficiency of OFDMA or the savings of the femtocells, will help to an extent, but by that fateful 2012 date, the operators will need to be thinking seriously about whether they want to be utilities or internet service providers.