Clients | Basics of Wi-Fi

A client is the typical end-user device. Unlike access points, which are strategically placed for coverage, clients are almost always mobile (or potentially so).

Wi-Fi clients can be general networking interface devices, such as those in laptops, or can be part of a purpose-built mobile voice handset. Either way, these clients appear to the network as endpoints, just as Ethernet devices do.

From the user's perspective, however, Wi-Fi clients add an extra complication. Unlike with wireline connections, where the user is assigned a port or cable and has the expectation that everything will work once the cable is plugged in and the process has settled down (which, for administrators, generally means that Dynamic Host Configuration Protocol (DHCP) automatic IP address discovery has completed), wireless connections have no one cable to solve all problems. The user must be involved in the connection process, even when the reason for connection or disconnection is not readily apparent. As mentioned previously, the user must learn about SSIDs. When a wireless interface is enabled, the user is normally interrupted with a list of the available networks to connect to. Knowing the right answer to this question requires an unfortunate amount of sophistication from the user, not because the user does not understand the technology, but because they usually do understand the power of mobility, and have learned to strategically hunt out wireless networks for casual email access. This is clearly evidenced by the pervasive nature of the "Free Public WiFi" ad hoc (Independent Basic Service Set, or IBSS) SSID that tends to be on so many laptops.

Ultimately, the user is responsible for knowing what the appropriate network is to connect to at any given location. Most devices do remember previous connections—including authentication credentials, in many cases—and can make the connection appear to be automatic. However, because of that caching, installations that run multiple SSIDs are often forced to deal with users not knowing exactly which network they are connected to.

Once the connection is established, the interface comes up much as a plugged-in Ethernet link does. Any automatic services, such as DHCP or Universal Plug and Play (UPnP), that run on interface startup will get kicked off, and the users will be able to communicate as if they had plugged directly into the network.

The last wrinkle comes, however, with mobility. Once the user leaves the coverage range of the one access point that it is on, the client will perform its list gathering activity (scanning) again. If it can find an SSID that it already has in its list—especially if the SSID is the same as the one the client was already associated to—the client will try to hand over to the new access point without user intervention. However, if the handoff does not succeed, or there are no more known networks in range, the client will disconnect and either warn the user with a popup or just break the connection without warning. This can come as quite a shock to the user, and can lend negative impressions about the network.