Your browser is not supported. Please update it.

18 December 2007

New markets make femtocell economics more attractive than homezones

The main approach to fixed/mobile convergence to date has been combining dual-mode Wi-Fi/3G handsets with an in-home access point, backhauled by the user’s DSL line, to create a homezone. The 3G femtocell has been lurking on the horizon all year, offering a potential alternative to this somewhat clumsy dual route, and one that gives operators greater control of their user base and revenue streams. Trials and RFPs are sprouting around the world, and the technology is looking increasingly usable ‘ provided the price points make sense, both for the operators and vendors. With majors like Vodafone and India’s BSNL setting aggressive expectations on pricing, even at the pre-mass market stage, some equipment suppliers are questioning whether this will support a viable business. In particular Ericsson ‘ with an obvious interest in slowing any tide towards commodity base stations ‘ and Texas Instruments have been expressing their doubts, even as the surge of carrier interest suggests they will not be able to stay out of the trend for much longer. In the short term, hope may lie in the developing economies, where operators with little existing 3G build-out are looking to femtocells to support a new way of deploying next generation RANs, one that could impact their business cases significantly even at higher price points.

Examples of this include Vivo in Brazil (whose leading landline operator Brasil Telecom is also demonstrating high interest in 802.16e femtocells, if it gets significant WiMAX spectrum), BSNL itself and JSC Ukrtelecom of Ukraine, the latest in a line of carriers owning up to femtocell interest at various conferences held in the past few weeks. All these have in common that they have either not started 3G deployment, or have very made very little progress so far, and so have a relatively clean slate in deciding on their technologies and network designs. In addition, all are looking to deploy a combination of 3G and broadband wireless, to attempt to support a more profitable model than 3G alone, and to leapfrog the more developed countries in infrastructure terms. Key aims within this include rapid build-out of urban or high income population zones, to achieve short term profitability by delivering very high performance (a luxury denied to the European 3G cellcos); and the need to support superior indoor penetration in the absence of extensive DSL or cable build-out. For such operators, the flexibility of a RAN that is heavily designed around femtocells can represent something far more disruptive than merely extending the network indoors at low cost and creating homezone structures (the key first phase objectives of the western femtocell supporters). Ukrtelecom only launched UMTS services last month, and so is looking to femtocells to accelerate its roll-out to key target user bases while improving quality of service and so increasing ARPU. The incumbent is trialling products from Huawei and Nokia Siemens.

In turn, this more visionary approach to femtocells ‘ as a ‘bottom up’ entry to next generation networks and convergence ‘ may support different economics, with greater margin potential for both operators and vendors, or failing that, massive volume (especially if BSNL’s plans translate into dollars) and the chance for suppliers to take an early stake in a developing market, possibly without quite the level of bureaucracy and risk that have dogged recent major infrastructure awards in India and elsewhere, forcing some suppliers to withdraw altogether. In the 4G stage, the need for more flexible approaches to deploying RANs will become vital to the operator’s ability to make a profit, and while Ericsson may be clinging to the last vestiges of its high margin, proprietary macro station mode, this is dying quickly even without the help of femtocells. Nowhere is this more obvious than in India, where Nokia Siemens and even Ericsson itself have withdrawn from a portion of BSNL’s vast GSM expansion deal because of the razor thin margins this would entail.

However, in the developed economies, the 3G operators have far more investment in their networks and, especially in Europe, an urgent need to generate some form of return on that, even in the face of consumer demand for flat rate mobile broadband and unfettered internet access. Femtocells can help them, as outlined above, but only if they are at very low cost ‘ a point Vodafone CEO Arun Sarin very publicly defined as ‘substantially less than $100’ per unit (see Wireless Watch July 23 2007), and which other carriers are now pushing even lower.

BT added its weight to the argument last week at a femtocell conference in London, where it talked about its ongoing technical evaluations of the miniature base stations, but also sent a clear signal to potential suppliers that it still considered the technology too expensive when compared to Wi-Fi and dual-mode. ‘We’re doing technical evaluations of femtocells, like everyone else, but the big issue is cost,” said Warren Buckley, BT’s director of mobility and convergence, speaking at the Informa Femotcell Solution conference. “The cost of the BT Home Hub is significantly less than 50% of what we’re being quoted for a femtocell. The cost model now does not justify me putting a femtocell in the Home Hub.”

BT was the first major adopter of a dual-mode convergence approach based around UMA (Unlicensed Mobile Access), and while UMA can also be used for femtocells, its comments reinforce the notion that the tide is turning against dual-mode and in favor of femtocells (and, if the US is a bellwether, SIP rather than UMA or other mechanisms). The main cost-related point in femtocells’ favor is that there is no need to support dual-mode or otherwise specialized handsets, a factor that has held back BT’s Fusion converged service, whose uptake in the UK so far stands, reportedly, at fewer than 50,000 subscribers after two years. This also seems to be influencing Orange and Telecom Italia, both of which were early UMA/dual-mode adopters but are now in serious femtocell trials. Telecom Italia has reportedly relaunched its UMA-based convergence offering with a SIP-based system, and will introduce femtocells ‘at an early stage’, according to an insider. Both the incumbent and Vodafone Italy have run into regulatory difficulties with UMA as well as finding a shortage of handsets. In the short term, Telecom Italia’s new version of its Unica offering still uses dual-mode devices, although SIP does not support handover between Wi-Fi and cellular and so the convergence is achieved through a specially developed client running on Symbian handsets. This is driving the telco to an upgrade to a system based on IMS and femtocells, achieving fixed-mobile convergence without Wi-Fi.

While Ericsson’s attempts to pour cold water on 3G femtocells are understandable, Texas Instruments’ scepticism is perhaps to be taken more seriously, especially ‘ from a volume point of view – if it keeps a major chipmaker out of the market for any length of time.
The company’s infrastructure division says that, despite finding the technology interesting (and sponsoring one of the recent femtocell conferences) it has so far failed to find a convincing profit model, and is actually considering transferring the activity to the handset operation, which has greater experience of creating chips for consumer products. In the mean time, the hesitancy of TI, Qualcomm and Broadcom, and the exit of STMicro, creates a major opportunity for the chip designers that have specialized in this market at an early stage, notably the UK’s picoChip (with some activity from ADI, Xilinx and, in WiMAX, Sequans). While some would-be femtocell players, such as RadioFrame, Airvana (in CDMA) and NextWave, are designing their own silicon, picoChip has dominated the nascent market so far, and is the only company to have shipped a baseband specifically designed for a femtocell at this point. The company’s PC6532 reference design delivers an integrated PHY and MAC Wave 2 solution in a single chip, the picoArray-based PC205 processor. The reference design supports WiMAX Wave 2 and full IO-MIMO with Matrix A and Matrix B downlink, and both MRC and UL-CSM uplink. The WiMAX architecture is scalable from femtocells to multi-sector macrocells. The PC205 processor integrates an ARM processor for MAC functionality, and picoArray multicore DSP for software defined PHY.

Femtocell forecasts:

France-based research firm iDate expects limited deployment of femtocells in 2008 because of standards, interoperability and outstanding technical issues, but expects these to be largely resolved by 2009 (also the year Ericsson sees 3G femtocells become a reality). iDate predicts this will lead to shipments of 10m UMTS units in 2010 and 18m in 2011.

In a recent forecast, ABI said 74m femtocells would be shipped between now and 2012, half of these between 2011 and 2012.

UK-based Analysys, which specializes in operator business models, is also upbeat about femtocells, saying in a new report: “The idea of femtocells is a really clever way to avoid needing to make these big 3G networks more dense to provide the necessary quality of service. Some operators would have to double or triple the number of base stations they have to provide the very high quality of service required for indoor coverage.’ However, it will be vital to balance the decline in revenue from premium voice tariffs with new revenue generating applications supported by femtocells and homezones, such as location-based services. If the only business objective is to support low cost calls in the homezone, the financial impact could be ‘disastrous’.

In its rearguard action against 3G femtocells, Ericsson is more positive about GSM versions, and offers such a product itself already. Some operators are showing interest in these devices, most recently France’s Bouygues Telecom, which is using Nokia Siemens products. However, in most cases these 2G deployments are more akin to classic indoor picocell systems for enterprises, shopping malls and so on, rather than home-based consumer roll-outs, with specialists like RadioFrame (which partners with NSN) apparently adopting the femtocell label for these more traditional devices for marketing purposes.