The number of mobile broadband subscribers (using W-CMDA, EV-DO, HSPA, LTE, and Mobile Wimax) is expected to surpass two billion by 2013, with LTE alone reaching 440 million subscribers by 2015.
This growing demand for next-generation wireless broadband introduces a new level of complexity. 4G networks are intended to provide users with data rates of more than 100Mbit/s.
Compared to current W-CDMA networks that support data rates of 2Mbit/s with additional rates improvements for HSPA, 4G modems will require almost a twenty-fold increase in processing horsepower.
In spite of the significant increase in complexity, handset OEMs are still allocating the same power budget for 4G designs in order to fit into battery-powered devices.
Designing a powerful and cost-effective communications IC that meets these stringent power constraints poses a fundamental challenge to the SoC designers.
Routes to 4G baseband
Although the term 4G is widely used to describe the next step in wireless communication, we are still lacking a precise definition of the architecture and its deployment path.
Analysing the routes to 4G wireless baseband, there are two evolving technologies that are competing for the leading position: LTE and Wimax.
Wimax has emerged as the technology of choice for computing devices such as notebooks, netbooks and MIDs. It is heavily promoted by the Wimax Forum, led by Intel and other companies. The deployment of Wimax has already begun but its success is far from being secured.
LTE is evolving as the “GSM” direction towards cellular broadband and is promoted by the 3rd Generation Partnership Project (3GPP). Consequently, it appears to be the preferred choice for mobile handsets and smartphones as it is widely promoted by baseband OEMs and operators including: Qualcomm, Ericsson, Verizon, Vodafone and others.
It is unclear which 4G standard will prevail, and it is possible that both standards will coexist and serve different uses in different geographies.
The changing standards increase risk with a traditional, hardware-based design approach. IC designers using such an approach can either find themselves betting on the wrong standard and end up with an obsolete product before even being launched, or getting into costly investment by developing dedicated chips for each of the two standards, with no guarantee of a return on investment.
To ensure that both standards are supported in 4G modems, a flexible approach is preferred, and this means a programmable engine capable of delivering the required performance.
Multiple interfaces
The evolution of mobile modems is not restricted to the complexity increase of wireless standards. Today’s smartphones already need to support multiple wireless air interfaces. 4G mobile devices will need to support, alongside Wimax and LTE, a large number of wireless air interfaces, including GSM, GPRS, EDGE,
W-CMDA, HSPA, Wi-Fi, Mobile TV (DVB-H) and Bluetooth.
Traditional hardwired solutions require multiple dedicated baseband modules, each targeting a different standard. For addressing multiple standards, such hardwired solutions require the duplication of large silicon areas, making them very costly. Hence such an approach cannot economically keep pace with the growing number of evolving standards.
A software-based approach that offers multimode systems support and can replace multiple dedicated baseband systems seems to be the right direction. In addition, this rapid evolution requires a more efficient development process. Reusing the same platform over multiple product generations is essential.
A programmable approach enables the required reusability and ensures fast time-to-market.
As we move to 4G, the cost of development and the changing standards increase the risk of a traditional, hardware-based design approach. This naturally leads to looking at a programmable multimode solution. Still, there is a huge challenge in providing a programmable system that is capable of handling the tremendous processing demands of 4G standards under stringent power consumption constraints.
Standard programmable architectures cannot efficiently address these challenges. The best approach is a specialised communications processor that is optimised for the most demanding wireless standards.
Such a communication-specific programmable approach can address these performance and flexibility challenges and enable efficient 4G modem designs.
Eyal Bergman is director of product marketing at Ceva.