Gideon Wertheizer, CEO of CEVA - a licensor of silicon IP platform solutions and DSP cores - talks to Electronics Weekly about mobile handset design, the crossover with netbooks and the true potential of software-defined radio...
1. Describe CEVA's business & technology in 2 sentences.
CEVA is a leading licensor of silicon IP solutions and DSP cores for mobile handset, consumer electronics and storage applications. Our technology is licensed to leading semiconductor vendors, which in turn manufacture, market and sell ICs based on CEVA technologies to systems companies for incorporation into a wide variety of end products.
2. How is the approach to mobile handset design changing?
Over the past few years, changes in mobile handset design have been driven by market evolutions in two different segments: the ultra low cost phones serving the need for sub $20 phones in regions such as China , India and Latin America, and the high end / smartphone markets where wireless communications (e.g. 3.5G), computing and multimedia have converged into unified platforms.
For ultra low cost handsets, power consumption and the overall Bill of Materials are the critical success factors of a platform design. For smartphones, performance and feature sets are the decisive factors and require the integration of connectivity and multimedia technologies such as Wifi, GPS , Bluetooth and High Definition video and audio into the design.
Handset designs for these different end markets require different levels of hardware integration and software. And in order to deliver the final optimal product, each handset design should be further customized for geographical regions and even per operator.
All of these factors relate to a substantially higher design cost to semiconductor companies in the form of leading geometrics design, design of advanced wireless standards such HSPA+ and LTE, integration of connectivity features as well as multimedia and 3D graphics, all under constraints of low power consumption.
As a result, of these factors, it has become impossible for any single player in the market to specialize in all these different hardware and software technologies and this has changed the approach for mobile handset design for the future.
Going forward, mass market handset design will come about through a collaboration model, where there will be a number of vendors whose solutions will be widely deployed across the handset OEM chain and help reduce development costs and improve functionality of these handsets. These vendors and technologies will include:
• Open source software vendors providing the complex Operating Systems (e.g. Google 's Android, Symbian).
• IP vendors providing embedded processors, multimedia, graphics and memories (e.g. ARM, CEVA , Imagination, Virage Logic).
• Fabrication houses designing platforms with all the common denominator technologies included.
This will leave the OEMs and semiconductor companies to undertake the branding, ecosystems, user interface and customizations of the handsets, which are the true differentiating features.
3. How is an IP-based approach changing the cost/performance of handsets?
As the design of handsets becomes increasingly complex, the ability for semiconductor companies to develop and maintain every block within a handset chip in-house has become a costly and resource intensive task. Today, most companies understand that there are critical blocks within the handset design that can be licensed from leading IP companies like CEVA and ARM that will deliver the optimal solution for that block of the chip, and at a lower price than they can develop the technology in-house. The DSP in cellular baseband is one of these components and customers who license our DSPs for their handset designs can save as much as 20-30% of the overall design cost for that chip, and with a lower power solution than anything they could design in-house.
This leaves our customers with more time and more budget to concentrate on other areas of the handset design, for instance, integrating more technologies onto a single chip or differentiating their solution through software, and delivering a better overall solution at the end of the day.
4. Is this approach being implemented in netbooks as well as smartphones?
In netbooks, there are a number of different chips that make up the design. From a connectivity perspective, every netbook will require mobile broadband connectivity in order to have the anywhere, anytime access that a netbook is designed to provide. The leading players developing the mobile broadband modules also realize that licensing IP for critical blocks of their modem design is the best way to achieve the optimal solution and at a lower cost.
From an applications / CPU perspective, we believe the same scenario applies - blocks of IP for video, audio, graphics and more that have been designed by IP players targeting battery-powered devices like smartphones are an excellent fit for the design of netbook systems and reduce the cost and design complexity of these devices.
5. What is the true potential of software-defined radio?
The industry has been dealing with software-defined radio (SDR) for more than a decade, but has been skeptical about it as a legitimate technology for mass market portable devices. The technology barrier for running the complete wireless transceiver in software in a power-efficient and cost-effective manner was deemed too high. However, recently there is a growing interest in a software-based approach for wireless communication. There are a number of factors that have changed the feasibility of SDR:
1. The introduction of 4G wireless standards (LTE, WiMAX) means more processing power is required for performing the PHY processing
2. Flexibility to support growing number of wireless standards (LTE, WiMAX, HSPA+, HSPA, CDMA2000, EDGE etc)
3. Shrinking geometries means that more processing power is now available in the same die area; thus, cost-wise, SDR is now becoming a reasonable approach
4. Shrinking geometries play an important role in reducing power consumption to a point where power efficient SDR solutions become a reality.
As a result of these factors, a SDR architectural approach for next generation wireless infrastructure and handsets makes sense and will form the basis for the future of wireless processors. To meet this requirement for SDR processing, we developed CEVA-XC, an advanced wireless communications processor, supporting multiple air interfaces including LTE, WiMAX, HSPA, alongside 3G and 2G in software. CEVA-XC addresses the precise requirements of any 4G baseband design, from handset terminals and mobile broadband modules through to wireless infrastructure equipment.
See also: Q5 - Interviews with electronics industry leaders
Read all the Electronics Weekly Q5 interviews. From ARM's chairman, Sir Robin Saxby, to touchscreen technology firm Zytronic's MD, Mark Cambridge, the business leaders share their particular insights on the UK electronics industry.