Two years ago, I added wireless capabilities to my network via a 3Com wireless access point at an upgrade cost of around $2,500. Today, you can buy a wireless access point that does a lot more than that 3Com device for under $200. Now that wireless networking has finally gone mainstream, there’s little reason not to add wireless capabilities to your network. In this Drill Down, I’ll explain what you need to know about installing a wireless network.
Wireless networking hardware
Setting up a wireless network is similar to setting up a wired network. The biggest difference is the required hardware. In this section, I’ll introduce you to some of the most commonly used wireless devices and explain what functions each performs.
Wireless access point
The central component of most wireless networks is a wireless access point. Wireless access points come in all shapes and sizes, but you can see my older 3Com model shown in Figure A.
|This is a 3Com wireless access point.|
A wireless access point is a device that connects a wireless network to a wired network. The wireless access point acts as a hub for the wireless clients. It also contains a standard Ethernet port for connection to the wired network. This allows two-way communication between the two networks.
Although using an access point is the most common method for installing a wireless network, a wireless access point isn’t a requirement. You could use wireless network cards, which have two modes of operation: infrastructure and ad hoc. When running in ad hoc mode, the cards can communicate with each other directly, without the need for an access point. But using an access point makes the network more manageable, allows communications with the wired network, and gives you greater control over security.
Wireless broadband gateway
There are a million different brands and models of wireless broadband gateways (for an example, see Figure B) on the market with all sorts of different features. What they all have in common, however, is that they attach directly to your DSL or cable modem and share that broadband connection with wireless clients through a built-in wireless access point. Most models also include a small built-in hub for attaching wired clients as well. Likewise, most of these units have firewalls to help protect your network from Internet-based intruders. The cool thing about wireless broadband gateways is that most of them offer features that were only available to large corporations a couple of years ago. Generally, such products range from $200 to $600, depending on the features offered by the unit.
|This is an example of a wireless broadband gateway.|
Wireless PCI cards
Laptops use PCMCIA expansion cards, while desktop machines tend to rely on PCI cards. Fortunately, wireless network cards are available in both flavors. Figure C shows a wireless PCI card made by Linksys. Figure D shows a PCMCIA card. The black piece on the end of the card is the card’s antenna. Both cards operate at the same 11 Mbps speed, but are intended for different types of machines.
|Wireless PCI card|
Wireless USB NIC
Another type of wireless NIC is a wireless USB NIC. I’m especially fond of USB NICs because they will work on both desktops and on laptops. I’ve also run into situations in the past in which a PC will be low on IRQs, base memory addresses, etc., and I’ve been unable to make a wireless PCI card work. In such cases, a USB-based card works every time. You can see an example of a wireless USB NIC in Figure E.
|Wireless USB NIC|
Wireless Ethernet bridge
A wireless Ethernet bridge provides a way of connecting a wireless and a wired network together. While a wireless access point provides a way for wireless clients to attach to the wired network (and vice-versa), a wireless Ethernet bridge enables wired devices to function on a wireless network.
For example, one of my laser printers has a built-in JetDirect card, which allows it to plug into the network. I wanted the printer to work on a wireless network, but no wireless cards were available at the time. My solution was to plug the printer’s network card into the RJ-45 port on the wireless Ethernet bridge. In this situation, the printer maintains its own IP address, as does the bridge. When clients need to access the printer, the routing tables direct them through the wireless access point to the wireless bridge, and then on to the printer. In this scenario, I am using the wireless Ethernet bridge to attach a single device to a wireless network, but it is possible to attach an entire segment to the wireless network through the bridge. If you’re planning on attaching multiple devices, however, it’s cheaper and more efficient to use a wireless access point rather than a wireless Ethernet bridge. You can see an example of a wireless Ethernet bridge in Figure F.
|Wireless Ethernet bridge|
Wiring for wireless
Normally, when you create a wireless network, you begin the process by attaching a wireless access point to your wired network via a standard patch cable connection to your hub. Once the wired connection to the access point has been established, you must use one of the clients on the wired network to configure the access point.
Web interface issues
Most wireless access points can be configured through a Web interface. The units have built-in Web servers that host a configuration Web site. Likewise, the units also have a built-in DHCP server that can distribute IP addresses to wireless clients. If your network already has a DHCP server, you should disable the wireless access point’s DHCP server to prevent it from passing out IP addresses that have already been leased by another DHCP server.
Attaching to the wireless access point is as simple as opening Internet Explorer and entering the wireless access point’s IP address. You’ll have to look in the documentation that comes with your access point to see what IP address is used, but it’s very common to use 192.168.0.1. Initially, it can be difficult to attach to the wireless access point’s configuration Web site. If you have trouble, make sure that Internet Explorer isn’t configured to use a proxy server. If your network is dependent on a proxy server, add the wireless access point’s IP address to the proxy server’s Local Address Table (LAT), and you should be able to access the configuration Web site without having to uninstall the workstation’s proxy client.
You may also encounter a subnet mismatch. For example, if your home network uses a subnet mask of 255.255.0.0 and your wireless access point uses a subnet mask of 255.255.255.0, your network won’t be able to communicate with the wireless access point. This means that you may place everything onto a common subnet or update your routing tables to provide a logical path to the wireless access point.
Once you have made a connection to the wireless access point’s configuration Web site, it’s time to begin the configuration process. You must choose the settings you want to use, and then later configure your wireless clients to use identical settings. The actual configuration process for a wireless access point differs among the various manufacturers, but the basic information you must provide remains fairly consistent. In the sections below, I’ll discuss some of the more important wireless settings.
Wireless LAN service area
The wireless LAN service area, also called the SSID, is the wireless network identification. Usually, the wireless LAN service area is a text-based name. For example, I might call the wireless LAN service area something like "Brien’s Wireless." Such a unique name would virtually guarantee that I won’t accidentally use the same SSID as my neighbors.
A wireless access point has a long enough range to service most small offices. However, in larger office buildings, one wireless access point may not have enough range to provide all of the necessary coverage. In situations like this, multiple wireless access points can be used to get the job done. Thus, wireless network cards function in way similar to cell phones. As a user roams the building with a laptop, the wireless NIC looks to see which of the available access points is providing the strongest signal, and locks on to that access point until the signal becomes weak and another access point is needed. The unit name is the method the wireless NIC uses to determine which access point it is communicating with.
Although 802.11B wireless access points work on the 2.4-GHz frequency range, there are different channels within the 2.4-GHz spectrum. Generally speaking, there are nine different channels available, although some older models have only three to six channels. Selecting a wireless channel isn’t much different than tuning the radio in your car to a specific station.
Why so many channels? One reason is so that you may find a channel that isn’t affected by environmental factors. For example, when set to channel 3, my cordless phone affects network performance. Another reason for different channels is privacy. Imagine that you are operating on channel 6, and your office neighbor decides to operate a wireless network on channel 6 as well. Normally, it wouldn’t be a problem unless both of you chose to use the default SSID. If that happened, the two networks would likely interfere with each other.
Using a different channel than your neighbor is also a good idea from a performance standpoint. Like copper cable, each channel has a limited amount of bandwidth. Once a certain number of PCs are using a channel, performance tends to suffer. Usually, you have to have about 64 PCs on a channel before performance lags, but if a client is using the connection heavily, performance degradations can occur with only a few PCs online.
Wireless Encryption Protocol (WEP) is an encryption technology that prevents someone who may be eavesdropping on your channel from stealing packets out of the air and using them to gain access to sensitive information. WEP comes in a few different flavors: the 40 bit (now nearly extinct), 64 bit, and 128 bit.
Enabling WEP is usually a matter of simply selecting 64-bit or 128-bit encryption and then choosing a WEP pass phrase. The pass phrase is a 13-character (64 bit) or a 26-character (128 bit) hexadecimal number. The idea is that the number is never transmitted. Instead, the number is hard coded into the wireless access point’s and the client’s configuration. When a client sends a message to the access point, the packet is encrypted using the WEP pass phrase as the key. When the access point receives the message, it can decrypt it because it is already in possession of the necessary key. There is little danger of someone on the outside decrypting the message because the key (WEP pass phrase) is never transmitted, but rather is preshared.
The downside of wireless networking
An 802.11B network card is rated for a speed of 11 Mbps. Obviously, 11 Mbps is slow compared to 100 Mbps Ethernet. Even so, in the real world, I’ve never once seen an 802.11B network card actually achieve an 11 Mbps connection. 802.11B devices are designed to slow down if they are unable to reliably offer a higher-speed connection. Under ideal conditions, an 11 Mbps network card usually offers about 5 Mbps of throughput. However, under unfavorable conditions (such as a lightning storm), these cards can run as slowly as 1 Mbps, or stop working completely.
In spite of the problems associated with wireless networking, I’ve found that it makes my life considerably easier, and I can’t imagine going back to a completely wired network. As my network grows, I just make sure that anything that depends heavily on high network performance or reliability is placed on a wired segment. I only use wireless links for computers that either need mobility or are in areas that I can’t reach with a network cable.