The cellular communications industry is currently abuzz with two big issues: the development of Long-Term Evolution (LTE) and the emergence of femtocells. Although the two may at first sight appear barely related, it is becoming increasingly clear that they are, in fact, inextricably linked.

Even femtocell advocates sometimes underestimate the closeness of this symbiosis. They assume that the deployment of LTE femtocells will follow the model established with 3G, filling in for poor indoor coverage. But more informed voices now insist that femtocells will lead, not lag, the rollout of LTE.

LTE On The Horizon

LTE is the latest technology in an evolution that, in barely 10 years, has led from the “original” 3G UMTS specification, through high-speed downlink packet access (HSDPA), high-speed uplink packet access (HSUPA), and more recently to HSPA+. Closely linked with the emerging concept of the next-generation mobile network (NGMN), LTE has already attracted the backing of operators with existing networks based on a wide range of technologies.

LTE is designed to provide improved capacity, lower cost per bit, reduced latency, and easier integration with packet-based networks that use Internet protocol (IP). Its headline data rates are in excess of 300 Mbits/s. For operators, the goal is a system that can be profitably used to target premium services at users who need and can afford them, as well as to relieve capacity problems in areas where existing networks are congested.

However, realizing practical LTE networks presents significant technical challenges. LTE uses wider channel bandwidths, advanced coding, and orthogonal frequency-division multiple-access (OFDMA) modulation methods that require unprecedented signal processing power.

For best performance, the system needs to achieve high modulation density. (The system communicates many bits per symbol.) Also included in the emerging standard are techniques such as multiple-input multiple-output (MIMO) that combine signals from several antennas to enable more effective communication.

So far, so familiar: infrastructure and terminal makers are well used to implementing standards and protocols that stretch their technological capabilities. But LTE presents a further set of problems. Primarily, it pushes against many of the conventional boundaries of cellular telephony, restrictions that are often seen as the “laws of nature” of communications design.

At the same time it must meet head-on many of the very practical challenges that 3G operators have so far been able to work around or gloss over. Most notable of these are the poor coverage and low data rates experienced by many users indoors—an increasing problem now that more than half of mobile minutes are clocked up at home or at work.

These challenges—some new, some old—require femtocells (or at least small-cells) for their solution. What seems likely is a fundamental shift in the architecture of the network, with smaller cells, closer to the user, playing a key role in LTE deployment from the outset. So what can small cells achieve that large cells can’t?

The first answer is that they can deliver what femtocells were first conceived to provide—better indoor coverage. This need first showed up in 3G networks, which use quadrature phase-shift keying (QPSK, a modulation technique that transmits four digital bits of information per symbol) at a 1.8-GHz transmission frequency.

Operators rapidly discovered that this combination travels very poorly through walls, which is a significant problem given that poor indoor coverage is a major source of customer churn. The data-centric nature of 3G meant that they could not take the traditional approach used in voice-based networks of deploying ever-larger macrocell basestations with sufficient transmit power to overcome losses through walls. The femtocell was born as a way of eliminating this problem by putting the basestation on the same side of the wall as the user.

For LTE, this physical problem will be even worse (Fig. 1). Attenuation increases with frequency, so 2.5-GHz transmissions will suffer even more than those of 3G. And, LTE relies for its high data rates on 16QAM and 64QAM (16- and 64-level quadrature amplitude modulation), schemes that can transmit 16 and 64 digital levels (often expressed as four or six bits per symbol). Delivering 64QAM through walls is highly unlikely. Even 16QAM is borderline.

But without this modulation depth, LTE will deliver performance no better than existing 3G networks. Worse still, turning up macro basestation power to reach the indoor user increases the level of interference for other users and/or cells. The overall performance of the network is actually reduced.