December 12, 2000, 3:39 PM — Broadband wireless LANs are about to make their debut. Wi-LAN and Philips
Semiconductors just released the first application-specific integrated circuit (ASIC)
chip based on the high-speed 802.11a standard. Wi-LAN will integrate the chip into the
I.Will products it will deliver
this fall, and vendors such as href="http://www.radiata.com/index2.html">Radiata Communications and Clarity
Wireless are right on Wi-LAN's heels. Does this mean you should hold off for the
higher data rate products before implementing a wireless LAN? Let's take a closer
The IEEE finalized the initial 802.11 standard for wireless LANs in June
1997. The standard specifies a 2.4 GHz operating frequency using frequency hopping or
direct sequence modulation with data rates of 1 or 2 Mbps. In late 1999, the IEEE
published two supplements to the 802.11 standard: 802.11a and 802.11b. One of my href="http://mithras.itworld.com/articles/columns/net-geier-0323.html">previous
columns discussed the 11 Mbps direct sequence 802.11b standard, for which many
vendors have compliant products on the market.
The 802.11a standard is quite different from its 802.11b counterpart. Among
A wide variety of high data rates are available: 6, 9, 12, 18, 24, 36, 48,
and 54 Mbps. 6, 12, and 24 Mbps are mandatory for all products.
Operating frequencies fall into unlicensed national information structure (U-
NII) bands at 5.15 to 5.25 GHz, 5.25 to 5.35 GHz, and 5.725 to 5.825 GHz. This provides
plenty of bandwidth to support the higher speeds. In the US, the Code of Federal
Regulations, Title 47, Section 15.407, regulates these frequencies.
Maximum output power is 40 mW for products operating at 5.15 to 5.25 GHz, 200 mW
for 5.25 to 5.35 GHz, and 800 mW for 5.725 to 5.825 GHz.
The physical layer uses href="http://www.wilan.com/ofdm/main.html">orthogonal frequency division
multiplexing (OFDM) for modulating data before transmission. A combination of
binary and quadrature phase shift keying (BPSK and QPSK) provide the different data
Forward error correction (convolutional coding) compensates for an errored
packet without the need for retransmission. FEC reduces overhead when noise or
interference causes bit errors.