The quest for more bandwidth is never ending. Soon after vendors make one technical advance, they move on to the next. With 54 Mbps 802.11g wireless LANs now poised to gain a lion’s share of product shipments, vendors are examining ways to push the bandwidth plateau past the 100-Mbps mark.
Proprietary approaches are already available to reach that goal, but analysts anticipate that a standard approach, dubbed 802.11n, will soon emerge and then lead to product shipments in 2006.
A couple of factors are driving the need for higher-speed WLANs. First, more devices are being connected to these networks. In corporations, the number of mobile users is growing, as more employees are being outfitted with laptops and handheld devices.
At home, users are now connecting their PCs to TVs and audio systems, and moving digital music and video from device to device. Also, applications are becoming more complex, and voice and video are less tolerant of bandwidth fluctuations than are data transmissions.
Vendors, such as Atheros, Netgear and D-Link have been ahead of the higher-speed WLAN curve and developed proprietary techniques, dubbed Super G, to double the speed of wireless connections.
To achieve the faster speed, they rely on channel bonding, which combines two 54 Mbps channels into one 108 Mbps link. “To date, the Super G products have gained only limited acceptance and tend to be used for niche multimedia applications,” said Allen Nogee, a principal analyst with market research firm In-Stat/MDR, a division of Reed Elsevier.
A few factors hindered deployment. Because the products are proprietary, and therefore not able to interoperate easily, users have been reluctant to deploy the technology. “Because Super G product shipments have been low, vendors lack the economies of scale needed to drive down pricing,” noted Greg Collins, an industry analyst with market research firm Dell’Oro Group.
Also, the channel bonding technique can actually degrade bandwidth throughput in certain cases.
The most popular WLANs (802.11b and 802.11g) offer 11 channels — each channel functions as a distinct network connection for each user, so the more channels a WLAN offers, the more users a network supports — but only three channels are nonoverlapping, which means they can operate in the same area without interfering with one another.
Sometimes even when a Super G network combines two nonoverlapping channels, throughput problems arise on other links.
Last, higher-speed transmissions tend to be less effective as transmission ranges increase. If users stand within 100 feet of an access point, they may realize the 100-Mbps rate, but as they move further away, throughput rates may drop.
In September 2003, vendors began to examine ways to boost WLAN bandwidth in a standard fashion, and Super G is one of the potential techniques being considered. In July, Agere Systems, working with Atheros, Intel, Nokia and Sony, outlined a way to increase transmission speeds to 500 Mbps by taking advantage of two innovations.
The first is Multiple Inputs Multiple Outputs (MIMO), which uses two to four sets of antennas rather than the single set of antennae now deployed on each wireless connection.
The second innovation is increasing the size of each WLAN radio channel, from 20 MHz to 40 MHz, which is somewhat analogous to using a larger data pipe between each link.
Standard Being Developed
These two innovations create a data stream ranging from 125 Mbps to 500 Mbps on each link. This technique operates in the 5-GHz frequency range, which has less interference than the 2.4-GHz range in which 802.11b and 802.11g operate.
“We’ve seen a hesitancy from users to adopt 5-Ghz products (802.11a works in this frequency range), because of compatibility issues with their existing networks,” said Abner Germanow, an industry analyst with market research firm International Data Corp. Also, the Agere proposal relies on channel bonding to increase throughput and cuts the number of channels available in the 5-Ghz band in half from 22 to 11.
The Institute of Electrical and Electronics Engineers (IEEE) is at a nascent stage of developing the standard. The organization plans to finish collecting potential proposals by the middle of August. The next step, which is expected to take place in the fall, is to hear formal presentations from the groups backing different approaches.
After that, the 802.11n committee will determine which items to include and which to exclude.
Single Specification Sought
“The standards-making process is a high stakes undertaking where vendors try to gain competitive advantage by having their technology adopted as part of the specification,” IDC’s Germanow noted. As a result, a fair amount of posturing and debate is expected to occur during the process.
While there is a possibility that a few iterations of 802.11n will arise, the vendors would prefer to craft a single specification that mixes components from many proposals.
Because of the high stakes and the fact that squeezing more bandwidth over existing frequency ranges is technically challenging, the evaluation process is expected to take at least six months and could stretch out to a year, or more. After that stage is complete, vendors will need time to develop standard compliant products.
“Vendors are talking about delivering 802.11n products in the fall of 2005, but I think the more likely date will be the beginning of 2006,” In-Stat/MDR’s Nogee said. “That timing may actually work well because corporations and consumers should be starting to deploy at that time the multimedia applications that require the extra bandwidth,” he said.