In the dynamic wireless communications market, where multiple standards, technology convergence, market uncertainties and consolidation all occur on a regular basis, operators and the cellular infrastructure OEMs supporting them are looking for agility and economy. No single system design technique can address the differing use scenarios or the OEM's change in focus from hardware platform value to the range of managed services.
Previously, OEMs focused on available logic devices for their designs. Silicon suppliers took a device-centric orientation. "Build it and they will come," they said. Today, however, R&D resources have shifted away from systems manufacturers.
In the past three years, over $1 billion has been carved from OEM R&D budgets. All the major OEMs have made deep cuts, and the vertical integration model in systems design and development is gone forever. New efficiencies are in play and alternative problem solving needs to occur. But there's still a clear need to address many things simultaneously.
We're living in times that demand new thinking, new concepts, and new organizations all managed in new ways. A new, more efficient approach to R&D needs to take up the slack in the food chain.
Consolidation and slashed R&D
Add to this the fact that operators are adjusting to a new reality, as new strategies based on geographical and technological synergy are replacing global ones. Reorganizations and consolidation of operators and OEMs are in play with Verizon/MCI, Cingular/AT&T, Sprint/Nextel and Alcatel/Lucent with a portion of Nortel as most recent examples. (Expect more in the the next 18 months!)
This consolidation has a clear, long-term impact on capital expenditures in technology. Add to this dynamic the fast evolution of standards toward greater mobility and 4G. In the face of these rapid business and technology changes, OEMs are reducing their R&D expenses; two major players are making reductions of up to 25 percent within the next two years.
The next technology frontier is spatial processing. With high demand for spectrum comes the increased need for spectral efficiencies. Frequency (FDMA, OFDM, etc.) and time domain (encoding, etc.) are at their limits, so spatial processing has become the new focus. No single technique addresses the differing use scenarios. However, MIMO offers great promise with the promise of independent data streams across multiple antennas providing high data rates.
Figure 1: Increasing capacity requires spatial processing.
So what are the silicon solution options? The choice comes down to ASICs, DSPs or FPGAs. Microprocessors are too coarse-grained for cost-effectiveness. Managing the standards in flux with ASICs is not viable.
The desire for a reduced development effort makes DSPs and FPGAs attractive alternatives. The cost-reduction capabilities in highest volumes point toward ASICs and in slightly lower volume to FPGAs. Scalability for MIMO is good with ASICs and even better for FPGAs.
The communications systems vendor focus is changing. Until now, the traditional model has been to differentiate hardware by delivering the lowest cost-per-channel, while IP, applications and services all took a back seat. The emerging model turns this upside down with services rising to the highest value and hardware becoming commodity.
No single solution
However, some silicon suppliers claim they have it all and can do it all. No one believes it. There are a range of viable options for a silicon platform in communications systems, but no one company supports it as the problems are too great. The solution is a wireless ecosystem addressing the OEMs' needs and filling the voids with best-in-class solutions.
An ecosystem of expertise, technology and silicon addressing all digital logic needs is the way forward. The key question from the wireless systems OEM's perspective is "Does the ecosystem address all my pain points?"
The ecosystems partners, typically a generalist surrounded by specialists, can create best-in-class solutions, but executing this support infrastructure for OEMs requires deep insight in all current and future customer needs. When effectively executed, the support ecosystem partners allow the OEM to move up the value chain and resolve the essential change in delivering R&D support.
In contrast to earlier do-it-yourself days, an ecosystem created around an OEM's needs in the WiMAX space offers a range of R&D solutions. It provides designers with a choice of paths to system success, ranging from OFDMA blocks, FEC cores and OBSAI/CPRI IP to ready-for-use PHY, MAC and RF silicon solutions. These ecosystems provide key design building blocks that reduce the OEM's R&D investment and help accelerate system development.
A good example of a vendor ecosystem that can speed development and keep costs low has been created by Altera, Turbo Concept, Sequans and TelASIC. More information is available at www.altera.com/literature/presentations/pres-wimax_design_require_rf3.pdf
The communications infrastructure market is maturing. Triple-play demand is leading to consolidation. 3G+ and 4G require major architectural changes in the midst of system OEM reductions in capital expenditures, while at the same time they need to move up the value chain.
Silicon suppliers and their partners are stepping into the gap as advanced R&D providers to complete the R&D resource shift with ecosystems of expertise and a solutions portfolio.
About the author
Arun Iyengar is senior director of Altera's communications business unit in San Jose, CA.