* An In-Stat/MDR research report predicts the Ethernet switch market will more than double over the next four years, but a Canadian analyst predicts the trend towards wireless networking will reduce the demand for switches.
Scottsdale, Ariz.-based In-Stat predicts network equipment manufacturers
will ship Ethernet switches with a total of 502.8 million ports by 2008 — up from 193 million last year. A recent In-Stat report, Ethernet Switch Market Analysis: Gigabit Ethernet Getting An Edge, found manufacturers’ revenues dropped from US$13 billion in 2002 to $11.4 billion in 2003 due to the decline in average selling prices (ASPs), according to the report.
“”Clearly IT spending is starting to pick up, and that has impacted the switch market as well,”” said Sam Lucero, the In-Stat industry analyst who wrote the report. “”I think it was very much in line with expectations.””
But while Albert Daoust, director of special projects at Evans Research in Toronto, mostly agreed with the report’s projections for sustained growth, Iain Grant, managing director of the SeaBoard Group in Montreal, said Wi-Fi is effectively displacing Ethernet.
“”The brakes were put in big time in 2001,”” said Daoust. He added that in Canada, the Ethernet switch market was worth US$469 million last year and is on pace to reach US$569 million this year. “”The next thing is that most PCs are shipping the switchable gigabit NICs in them, so people have to add gigabit switches throughout the network to accommodate the demand.””
Meanwhile, Grant said wireless local-area networks (WLANs) using the Wi-Fi standard have been reducing network managers’ dependence on Ethernet switches.
“”Wi-Fi is replacing not only the yards of blue cable, but also the connectors. Wi-Fi is displacing that potential,”” said Grant, conceding that there’ll always be some need for Ethernet switches. “”Even though Wi-Fi could be displacing Ethernet cables and small Ethernet switches, certainly the larger aggregation of Ethernet will need switches.””
But Lucero said the declining selling prices will help encourage organizations to buy Ethernet switches. He added demand is bolstered by other factors, including the total number of end points connected to networks, an overall increase in the diversity of connected to those end points, an increasing diversity of traffic flow on the LAN, an increasing integration with wireless networks, a heightened security awareness, tighter integration with partner networks and a need to replace old equipment.
“”There is an increasing need to replace aging equipment,”” he said. “”A lot of businesses and educational institutions and government offices have either decreased or held off on their IT budgets over the last two to three years, and so we’re starting to see a response to the pent up demand that that created. There is the continuing drop in ASPs which is continuing to encourage purchases.””
According to Daoust, video telephony — which he said could hit its stride in about 2006 — is the technology to watch in the years ahead as far as sustained growth in the Ethernet switch market is concerned, but Lucero countered that IP telephony and W-LAN will probably be more significant than video telephony.
—Ian Palmer
* Canadian researchers have discovered nanotechnology that can be used to develop an optical network that’s 100 times faster than today’s equipment, but acknowledged there’s more work to be done before it’s useful to the private sector.
In a recent study published in Nano Letters, University of Toronto electrical engineering professor Ted Sargent and his colleagues discussed the ability of a laser to direct another one with the control needed for fibre optic networks.
Sargent said his work, the result of collaboration between U of T and Carleton University in Ottawa, aims to speed up Internet traffic by reducing the number of times a signal is converted between optical and electrical form.
“”You could have a direct optical connection from here to Vancouver or from here to Los Angeles … and it could go seamlessly in optical form,”” he said.
These findings will also allow researchers “”to make devices that can perform a trillion operations rather than a billion operations per second,”” he added.
Until now, researchers have been unsuccessful in acting on scientific predictions that light can control light, a failure known as the Kuzyk quantum gap, after Mark Kuzyk.
To disprove the Kuzyk quantum gap, researchers designed a material combining nanometre-sized spherical particles called buckyballs (molecules of carbon atoms that look like soccer balls) and polymer. This combination created a clear, smooth film that allowed light particles to pick up one another’s patterns and could therefore process data carried at telecommunications wavelengths.
It’s hard to predict how long it will take for this milestone in nanotechnology, the result of two years’ work, to translate to practical applications in the commercial sector, Sargent said.
“”There’s at least maybe two or three more things that you need to do in order to make an all-optical network,”” explained Sargent. The first is to turn the researchers’ polymer-buckyball material into switches, and the other is to build a network using these switches.
Although Sargent cannot speculate on the costs of transferring this research outside of the lab, he said the materials and the processes his research team developed are inexpensive.
One member of the nanotechnology industry who’s been following Sargent’s work described using photons instead of electrons to route data as a “”massive, up-and-coming area”” in which hundreds of research teams in academia, government labs and industry are working.
Making a mainstream industrial product in large production runs, for instance, requires many steps, such as defining a product, said Neil Gordon, Montreal-based president of the Canadian Nanobusiness Alliance.
“”The key is to find an early-stage niche opportunity that is going to be adopted by some high-end user, or let’s say high-risk user, like the military or NASA,”” Gordon said. “”They will pay a lot of money to develop a first-of-a-kind technology.””
After prototyping the technology, the next step is to implement it in “”larger scale applications,”” such as optical networking, PCs, medical devices or networks on submarines, Gordon said.
Still another challenge is finding a user who wants to define certain specifications using the technology, which is “”almost as difficult to do as the research itself,”” said Gordon.
He added another caveat: “”Just because something works tremendously in the lab doesn’t necessarily mean that it’s going to be adopted in business.””
—Fawzia Sheikh
