Software Radiohead

By virtue of their given names, and perhaps a few predisposed genes, some people seem destined for particular vocations. Young Clete Boyer probably always knew he would be a baseball player. Likewise for Cal Ripken; he had the name, the pedigree and the talent to fulfill his destiny. And could Dick Butkus have been anything but a middle linebacker?

Maybe Vanu Bose could have been a fashion designer. His name would look good in a Corsiva font sewn into the labels of gowns worn Oscar night or etched into bottles of perfume. But by virtue of his genes, Bose is happily predisposed to technology. His last name is likely imprinted on your sound system's speakers, or at least the ones you wish you had. However, his first name — as suitable to a high-tech start-up as Clete was for baseball — is the one the 37-year-old academic-turned-entrepreneur hopes people will soon recognize.

Vanu Inc. was founded in 1998 to bring to market the software radio technology Bose invented while earning his doctorate at MIT. He completed his thesis on the technology in April of 1999 and earned his Ph.D. in electrical engineering and computer science two months later. His radio revolution almost stopped there as he considered an academic career, perhaps joining his famous father Amar (who founded the audio design company bearing his surname in 1964) as a professor at MIT. But the younger Bose reconsidered, and an entrepreneur was born.

“When you have an idea and you want to see it out in the market, you have to take the responsibility and the initiative to drive it through,” Bose said. “Just coming up with the idea by itself isn't all that valuable.”

As CEO, the responsibility is now his, and he harbors no illusions about the challenges ahead. “I have a great respect for how hard the process is when taking a great idea from a prototype to a product that is usable in the marketplace,” Bose said. “The challenges in that process are as [difficult as] developing and creating the idea.”

Bose's idea is that his all-software approach to radio technology in handsets and network infrastructure will make deploying or supporting new wireless technologies and services as easy as, well, a software download. That approach could also help engineers find the Holy Grail of spectrum optimization: on-demand radio channel allocation.

After an initial $300,000 angel investment, the company has been self-funding thanks in part to some government contracts and a commercial deal with Boeing. Today, Vanu licenses radio applications, or waveforms, including GSM, GPRS and UMTS; sells signal processing software; and provides reference design and consulting to companies that build radio devices.

Building for tomorrow, the company continues to engage in fundamental research that it hopes will garner patents for future software-based radio technology. It is working with companies such as Intel on low-power processors that will one day help make the software handheld radio feasible. More near-term, Vanu is counting on its software-based radio base station to challenge the hardware status quo.

Vanu developed its software base station using a homegrown Radio Description Language (RDL) that waveform engineers use to construct the software radio application. The main modules in the base station, which is designed to support all waveforms a mobile operator may need now and in the future, are written in C ++.

When the future arrives for handset technology — and both handsets and base stations are software-based — users will be able to tune their devices to talk to any other device, as well as turn them into walkie-talkies, cordless phones for the home, wireless data devices or things as mundane as baby monitors, Bose said.

Besides consumers, emergency organizations can benefit from the software radio's ability to allow for immediate interoperability between disparate technologies. Using Sept. 11 as the ultimate emergency situation — as well as the Gulf War and the war in Grenada as two military examples in which lives were lost in part because personnel could not communicate — Bose said his software will allow radios to be configured on the fly and made compatible.

“Communication becomes an issue when events are unplanned and you don't know in advance who will arrive with what equipment,” he said.

Obviously, however, the biggest benefit of software radio technology will go to the network operators — otherwise, it wouldn't see the light of day. Clear advantages to a software solution include lower costs for upgrading new products and supporting new waveforms, the quick introduction of new services, lower maintenance and remote management. More enticing still is the potential for operators to centralize their base stations using fiber to connect to cell sites. This is good for centralized management, but it also enables dynamic radio channel allocation so operators can move spectrum around the network at different times of day to cover known and unknown peaks in traffic.

“It lets operators fluidly move processing resources to where demand is,” Bose said. “Combined with base station hoteling, this will radically change the way you think about building cellular networks.”

Software-defined radio needs to address problems operators are facing today — and it will, said Phil Marshal, analyst with The Yankee Group. “It's quite a heterogeneous environment today, and rather than building out dedicated infrastructure for each wireless technology and creating different handsets, it makes sense to create solutions that traverse the different technologies,” he said.

That logic hasn't escaped the established players in the equipment space with which Vanu will compete. Most are working on introducing more software-based technology into their base stations. “Over time, a greater portion of specific functions are being pushed down to software,” Marshal said. He described their approach to change as more evolutionary than the one proposed by Vanu. After all, he said, their embedded hardware base accounts for approximately 70% of the operational infrastructure expense in the network and is worth protecting.

Marshal also said there are several hidden costs when start-up companies support new technologies. In Bose's view, it is too early to worry about those costs. He knows the mass market for his technology is still a few years away and there are still mountains to climb before he gets that first serious look from a major operator.

“I don't think [the competition] yet is worried about us,” Bose said. “After a successful field trial, maybe. After an order from a customer, certainly. Then people will get worried.”