Make sure you're ready for 4G

4G technology stands to be the future standard of mobile wireless devices and the successor of 3G in mobile handsets. In general, a generation is defined by a technology breakthrough over a period of time. In this context, 4G refers to the development of a technology change around a new air interface that provides higher data rates, or around the way we transfer data from an end-to-end point of view. For our purposes, let's look at 4G as the technology breakthrough around a new air interface.

This definition could cause some confusion because there are two different overlapping ideas about 4G, and because no killer applications have been defined. If you accept that video, following audio, will be the next data stream transferred over the air, you might have identified the killer app that's going to require end users to switch from 3G to 4G.

One idea behind 4G is the development of a wireless link at 100 Mbits/s while moving and 1 Gbit/s when stationary. The speed capabilities are being pushed by the mobile operators in Asia/Pacific (notably NTT DoCoMo in Japan). Another idea revolves around the development of a generic wireless network clever enough to connect to several present and future wireless access technologies (WLAN and cellular technologies like UMTS, EDGE, etc.) and seamlessly move between them. This vision of 4G, as a seamless communication environment calling for interoperability, drives the need for software-defined radios (SDRs).

The first idea has the benefit of being application specific. It's easy to imagine products associated with this concept because it's an extension of what we already have using 3G, but with the increased potential of operating at specific "hot spots" elsewhere. 4G proponents are also quick to point out the 3G limitation to deliver video/TV and Internet access. WiMAX and DVB-H seem to be appropriate alternatives, but their air interfaces aren't a breakthrough. These systems are based on OFDM, not a new air interface.

An alternative is 3G-LTE (Long Term Evolution). To make an appropriate decision on 3G-LTE, the performance (throughput, capacity, latency, and cost) must be evaluated and the launch date should be compared with 4G. In this case, 4G is a completely new air interface with enhanced mobility features over and above 3G-LTE.

The second idea forces us to focus on four issues:

• deciding upon (or at least understanding) some philosophical ideas for the future. Supporting circuit switched and IP protocols at the same time might not be efficient or feasible to develop a flexible protocol. Multi-carrier modulation, although bandwidth efficient, don't necessarily implement easily compared to a single-carrier modulation solutions. To a certain extent, the discussion between the benefit of CDMA and OFDM will need to be understood and probably both supported.
• understanding the different wireless standards (3GPP/3GPP2, 802.16, 802.11, 802.15, etc.). The GSM core network today can provide more than 10 Mbits/s through HDSPA, while the CDMA core network provides up to 5 Mbits/s through 1xEV-DO. The WLAN network reaches 108 Mbits/s with some enhancements for quality of service (QoS) at 5 GHz and with OFDM and MIMO. All those protocols (including others not listed) will need to be supported and harmonized. Some of these rely on TDD, others on FDD, and some on both.
• identifying what are the disruptive technologies. In the past few years, IC makers have paid particular attention to advanced signal processing mechanism, MIMO, OFDM, error correction coding, MAC technology (data rate and QoS), RF technology (range and sensitivity), as well as re-usable IP blocks and communication architectures.
• predicting the future of silicon technology to enable 4G performance. This significantly relies on the gate density of future digital processes and their capacity to support RF and mixed signal efficiently and simultaneously. And this must be done while keeping with Moore's Law of reaching a new technology node every 18 months.

As stated, the definition of a new generation for wireless communication includes a new air interface. The technology components mentioned aren't necessarily new. They may be widely used, but they aren't new. This is one reason that it's difficult to define 4G, because the two ideas discussed don't necessarily rely on a breakthrough technology. Rather, they rely on a few technology components to be well understood, new communication architectures, and a high level of integration on silicon. This is at the same time good and opportunistic. It's good because the industry isn't adding technical risk, and it's opportunistic because the entire 4G concept relies on the semiconductor industry to invent a new and flexible communication IC architecture that can be cost-efficiently integrated on silicon with acceptable performance.

Some technology components required for 4G are well known or already developed but might need tuning and improvements. Interference management, forward error correction (FEC), mesh networking, and robust QoS are a few examples to consider. OFDM also becomes a key technology because 4G relies on the increased spectrum capacity within a given spectrum, frequency, code, and time multiplexing and/or diversity, as well as a multi-carrier modulation scheme.

Other technology components of interest for 4G have minimal use in today's systems or are just coming out as key technologies in some applications. Space-time coding and MIMO (beam forming) provide diversity, space-division multiplexing, and coding gains (such as LPDC) for link budgets searching for more than 10 dB of gain and range increases. It's interesting to note that these technologies benefit from multi-path diversity, making MIMO and OFDM key technologies to master.

4G communication platforms will require networking processing power, applications support, a MAC, digital baseband, and analog front-end, and an efficient and low-power LNA and PA, which is similar to what's required today. Hence, the main hurdles of 4G are:

• The key technology components mentioned aren't always integration-friendly from a cost point of view with silicon processes available today. MIMO, for example, requires twice the amount of RF transceivers as a traditional implementation.
• 4G communications ICs will, and are already requiring unprecedented digital, analog, and software integration. This reinforces the need for improvements around system and software engineering, IC design methodology for 65 nm and below CMOS geometries, and IC design for manufacture.
• The development cost of such systems would require the development or acquisition of software and silicon IP cores that can be reused to recover costs.
• The development of programmable DSP architecture (from the LNA and PA up to the MAC entity) that can cope with various modulation schemes.

About the author

Cedric Paillard is the vice-president of TECHinsights at Semiconductor Insights. He holds BS and MS degrees in VLSI System Engineering from the University of Manchester Institute of Science and Technology. He also has a Bachelor of Engineering from the Ecole Superieure de Technolgie Electrique. Paillard can be reached at cedricp@semiconductor.com.