Network Computingwww.networkcomputing.com

In the Beginning

Since the early 1990s, wireless LANs operating in all unlicensed spectra (900 MHz, 2.4 GHz and 5 GHz) have been available on a limited scale. However, because of price, performance and interoperability issues, they were purchased primarily for vertical applications. We tested these devices and didn't recommended them for very many situations.

In June 1997, the first of the 802.11 standards was approved. It specified the physical and MAC (Media Access Control) layers for interoperable wireless LANs. In the physical standard, options existed for DSSS (direct-sequence spread spectrum), FHSS (frequency-hopping spread spectrum) and diffuse infrared. All wireless devices transmitted at the basic 1-Mbps data rate, and nearly all could also transfer data at the optional 2-Mbps rate. Most vendors opted for the DSSS option--no one implemented infrared.

In July 1998, the IEEE proposed another specification, setting the ground- work for 802.11b. This specification moved forward with only the 2.4-GHz DSSS physical layer (using 11 channels in North America), because most vendors were on that path. The major improvement in the "b" revision is the addition of data transfer rates at 5.5 Mbps and 11 Mbps. Plans are under way for operating in the 5-GHz range and offering data rates up to 54 Mbps. But don't count on those rates any time soon: The dust hasn't settled from the introduction of the 11-Mbps units.

Also part of the 802.11 specification is WEP (Wired Equivalent Privacy), a method of encrypting data using the RC4 algorithm between the client and the access point. WEP not only offers privacy, but also adds authentication by requiring an encryption key for access. Support for WEP was spotty in the original implementation, but most 802.11b products include it. Just be forewarned that enabling WEP may decrease performance in some cases.

Wireless LANs are not terribly complicated to design. There are only two components--the access points and the clients. Any client can talk to any other device, including another client. It's just a matter of how you configure your software. In the most basic configuration, clients are put into ad hoc, or computer-to-computer, mode. No access points are involved. You can set up a network in your conference room in a matter of minutes and transfer files between laptops at 11 Mbps.

Once you are configured in ad hoc mode, all the computers in your area should be accessible in the same way they are with a wired LAN--provided they are all on the same channel. If you are having difficulty, keep in mind that some vendors, including Apple Computer, work in ad hoc mode on only a single, unchangeable channel. Ad hoc wireless LANs can be considered a temporary setup that will not get you access to other corporate network resources.

Widening Your Horizons

More commonly, wireless LANs are set up in infrastructure mode. Here, hardware-based access points are scattered throughout your organization and bridge the wireless users onto the existing wired network--think of access points as wireless hubs. Clients are PCs or other nodes that have been equipped with wireless LAN adapters.

During configuration of the access points, you enter an ESSID (Extended Service Set ID) to define the access point's network name. Your client machines need to be configured to associate with an access point with that ESSID. Some clients allow a wild card in the ESSID configuration. In this mode, they will associate with the strongest access point, regardless of its ESSID. In most cases, it is a good idea to configure the clients with the actual ESSID, rather than a wild card, to prevent them from associating with other access points that may be in the area.

If you have multiple access points on the same subnet of your network, you should give them all the same ESSID. A number of access points configured with the same ESSID form an ESS (Extended Service Set).

Most client packages can measure the signal strength of access points. You can use this information to properly distribute access points in your building. If you have a low signal but don't think you need another access point, a higher-gain antenna may suit you--that is, if your access point provides an external-antenna option.

With multiple access points, clients are free to move seamlessly between access points ("roam"), as long as the ESSID matches. This feature is built into the 802.11 specification. When a client starts losing the signal with its associated access point, it begins to search the area for a closer access point. Once a new access point is found, the client initiates an association with the new access point and a disassociation from the old one. By properly placing access points through your enterprise, clients can move about without losing access to the network.