June 6, 2006 VC Reveals How Network Technologies Get from Universities to Vendors to You Former wireless exec laments lack of basic research at companies. by Bob Brown - June 6, 2006 Noticing a drop-off in the number and impact of enterprise network start-ups in recent years, we've begun exploring network-oriented developments further down the pipeline, such as at universities and other labs. In recent weeks, we've spoken with a handful of industry players whose jobs include keeping track of university research that might make a difference in the market for networking and other products. Here's the first in a series of stories we plan to publish over the next couple of weeks. Redelfs has been with Foundation for a little more than two years, and his background includes being CEO at wireless semiconductor maker Atheros from 1999 to 2003 and spending more than 10 years at 3Com in a variety of sales, marketing and management positions. As a member of Foundation Capital's team, he tends to invest in and work with companies in the wireless field, such as Dust Networks and Azimuth Systems. Not surprisingly, Redelfs said a key to ferreting out technologies at universities that have commercial promise is building relationships with administrators and researchers at the schools. Atheros itself got funding from Foundation after Stanford President John Hennessey, then Dean of Engineering, introduced professor Teresa Meng to Foundation's Bill Elmore. Another Foundation investment, WAN optimization company Peribit Networks (now part of Juniper), also got its start at Stanford. A newer Foundation investment, SiBeam, go its start at the Berkeley Wireless Research Center, and Redelfs says the fact that one of its founders also helped to start up Atheros gave the venture firm the inside track in funding it. Universities vary widely in how sophisticated they are in handling technology licensing, Redelfs says. Leading tech schools such as Stanford and California, Berkeley deal so frequently with patent licensing requests that when you go in there, they pretty much just say "here's the book," he says. "They have a standard model and it makes it very easy to do these things." Some schools are skittish about losing good professors to start-ups, but others view such ventures as sabbaticals during which professors learn a lot and come back as even better educators and researchers, Redelfs says. He says it is also important for professors in technology to have some commercial experience, as many of the students they teach will wind up in the market and will benefit from knowledge of commercializing technologies. Whether professors or grad students are open to licensing their technology is another matter, though Redelfs says most are flexible. However, there are those who are concerned that commercializing their technology could limit the scope of what the technology could do. "What the consumer wants or what's viable in the marketplace may not be pushing the envelope on the technology," he says. "And some academics feel that is their role... not to be constrained by the marketplace but only by what the imagination can dream up." Redelfs says there is room in the industry for researchers who are open to commercialization and those who are not. "Sometimes their lack of constraints leads to the whole next generation technology, which doesn't have a market opportunity now but will in the future," he says. Basic research isn't getting done at most companies (he cited IBM as an exception), so it's important that universities continue to do it, Redelfs says. His short history of Peribit is that the company's technology stemmed from a Stanford researcher doing DNA pattern matching talking with his brother about extending such techniques to data compression. As for current themes among businesses being pitched to Foundation, Redelfs says: "If there's too much of a theme then it's probably over-invested and not a good investment... I struggle with the number of WiMAX companies and Ultra-Wideband companies. The guys trying to consolidate four or five cell phone chips into one. Too many guys are doing too many similar things." Security, virtualization and storage are the hottest areas for ideas, Redelfs says. Though in security, he says he has perceived a backlash against appliance companies. "Net managers are saying: There are too many point products that do one part of security, and I can't have a stack of appliances," he says. When asked about whether there is any push among corporate customers to diversify the components of their networks in an effort to secure them, Redelfs quips: "Diversity is good for venture capitalists, but I don't buy it. I think it's a marketing ploy by people to try to sell products that are in the minority out there." He says companies that go with three different operating systems for the sake of diversity will wind up with a management challenge that could outweigh any security benefits. "Now I just have to track three times as many patches and yeah, maybe there's a little less damage on the part of the network that goes down, but I think it makes a lot more work along the way." Read the article at NetworkWorld.com