It's every mobile subscriber's dream - one wireless handset that works everywhere. No more dropped calls due to inadequate network coverage and no more problems roaming onto other networks. Cellular phones and wireless modems will seamlessly switch modes to operate on whichever network is currently available for service and select the one with the strongest signal for the best throughput.
It sounds too good to be true, but the technology to enable this dream world is not far off. Various companies have been working for years to make the vision of SDR technology a reality. But practically implementing a solution is not just a matter of overcoming problems in design and engineering. Aside from technical aspects, issues such as fraud, billing of calls and levels of service will all be rearing their ugly head in the future.
Taking a brief look at the SDR market today reveals a slightly different picture than most people expected a few years back. The future for the technology is still there, but it has taken a few detours along the way. The ultimate dream described above is giving way to a solution that is more easily digested by carrier companies forced to deal with the new technology and all of its aspects.
A fork in the road
The most basic decision facing the SDR industry lies in where to implement software-defined solutions. It's both conceivable and possible to implement reconfigurable hardware into either mobile handsets or the base station radios - or both. So the question that it boils down to is - which is better, to have the base radio architecture adapt to the user or the user adapt to the infrastructure?
While the original concept of a software radio was most often promoted for the handset, many in the industry today agree that SDR on the network side makes more sense - at least initially. "I think first base station SDR has to be successful and then handsets will follow" says Raj Karamchedu, senior marketing manager for Chameleon Systems, a developer of reconfigurable communication processors. "The driver and the enabler is the capability on the base station side. There is no point in bringing reconfigurable handsets to market if no service provider can support that capability."
Verizon Wireless CEO Dennis Strigl also commented about SDR technology during a conference in London as he wrangled with questions about Verizon's uncertain approach to 3G. ``A handset that gives customers the capability for roaming worldwide may be the short-term solution, the very short-term. The long-term solution is to give our customers the capability in the network,'' he said.
As others do, Strigl evidently feels more inclined to provide flexibility on the network side, where it is invisible to the consumer. This is a rare but now more frequent departure from more common appeals for multi-mode phones and software-defined radios to adapt to existing infrastructure. This makes sense since adding SDR capability to handsets would raise the price and create significant challenges to maintaining the unit's expected battery life. Network side SDR solutions are more easily controlled by carriers and give more flexibility in offering a whole variety of services at different levels.
Interestingly, some companies have already offered SDR solutions for years. AirNet Communications has secured contracts with several GSM infrastructure providers for their AdaptaCell base station solution. This site equipment only needs a simple software upgrade to support both GPRS and EDGE protocols.
But no major carriers have been adopting software-defined base stations for their entire network yet - the early adopters are the smaller players or those looking to test the equipment in small portions of their network. This may change in the near future though as many major carriers are looking at flexible networks.
One problem, many solutions
There's several different approaches being proposed for implementing SDR technology into commercial wireless systems. Some new companies, such as Chameleon Systems and Morphics, are pushing ahead in developing totally new microprocessor architectures that are suited to dynamic reconfiguration.
But traditional players are busy integrating several radios on a single ASIC - in essence, building a "software selectable" radio. Also, other big players in the wireless market such as Xilinx and Altera are busy developing hybrid solutions in their programmable parts - ones that contain a fixed processing core with plenty of programmable logic to implement whatever type of radio the designer wants.
Each of these approaches is competing for attention in the market, citing their benefits or pointing out the drawbacks of competing architectures. The architecture that is closest to the true SDR solution are the ones that are completely changeable in real-time. These "morphing" processors that configure themselves on-the-fly are the ultimate solution, but some people question whether this type of flexibility is needed.
A handset that is capable of re-defining operational protocols within a time period short enough to go undetected by the user is generally the goal. In the case of voice, this time is less than about 20 milliseconds so it is undetectable in normal speech. But changing modes during a data session may have looser constraints depending upon the application. "Not all consumer wireless requirements mandate real-time reconfigurability. Even a tight restriction on real-time reconfigurability can be loosened artificially" adds Karamchedu. Many SDR solutions along the more traditional hardware path take a few hundred milliseconds to switch function, but this can be hidden fairly well in data applications.
Another option to a pure SDR handset is one that is simply multi-mode. This type of phone is actually "hardware defined" since the protocols are burned into ASICs, but software selectable. These are in existence today but in limited applications. While a CDMA/AMPS phone is currently feasible, one that supports a multitude of protocols such as CDMA, TDMA, GSM, GPRS, and WCDMA is not so simple.
Many of the major handset vendors such as Nokia are starting to see success in integrating multiple radios on a single chip. Qualcomm also has detailed plans to make GSM/CDMA chipsets that will couple their technology with the worldwide GSM community. Higher integration will lead to more options in this area, but they must still deal with the drawbacks of fixed hardware implementations. It's conceivable that SDR solutions in the network will complement multi-mode handsets in the future - so these paths may not be mutually exclusive.
A team effort
SDR technology is complicated by many aspects that are not technical in nature. Just because supporting multiple operating modes is technically possible does not mean that the various players are eager to adopt it. Since one of the broad goals of SDR technology is to enable incompatible systems to provide common services, collaboration between players is key.
Developing a wireless handset that is capable of accessing the networks of competing service providers throws a new element into the partnerships forged by carriers and handset makers. The primary service provider may be weary of relying on a competitor's network to provide adequate QoS to their customer at a reasonable cost. Roaming agreements between partners are one thing, but relying on your competition to help further your own business is an area that many companies have yet to tread.
Also, many tenets of SDR technology diverge drastically form traditional design techniques and methods used by radio designers over the years. If there's one thing that scares designers, it's unpredictability. Compared to the fixed solutions they're used to dealing with, SDR is, at best, organized chaos.
For this reason, the tools used to develop the more advanced SDR processing cores are very important considerations for designers. According to Karamchedu, "familiar development tools are a major hurdle in the development of the technology."
So while there has been substantial progress in developing SDR technology, many of the barriers to entry into the commercial market are still firmly in place. Slowly, these solutions will start to find their way into the consumer market, but any noticeable presence is still a few years out. We can certainly still dream though.
Dave Mock is a freelance writer covering mobile technologies and markets. He's published papers to educate investors in wireless markets that are available through Amazon.com and BarnesandNoble.com. He also speaks at seminars and provides training to corporate clients. His first hardcover book on investing in wireless will be published with McGraw-Hill in Spring 2002.