WLAN Components
The wireless network components can be roughly mapped to the OSI reference model and broken down into the following categories for the purpose of providing a degree of structure when considering various design issues:
Radio frequencies, transmit power, and antennae (layer 1, physical)
APs and wireless bridges (layer 2, data link)
Wireless routers (layer 3, network)
Wireless clients (higher layers)
Radio-wave frequencies, the transmit power levels of a wireless device, and the type and gain of antennae replace the cabling considerations at the physical layer. Similar to the LAN switches, APs aggregate wireless clients and bridges extend the reach of a wireless LAN at the data link layer. And wireless routers extend the wireless reach into the network layer.
All of this occurs in support of the wireless clients, which allow people to satisfy their incessant desire to communicate with a greater degree of flexibility, mobility, and convenience. Now, consider the generic characteristics of these WLAN component categories and the relevant issues that must be taken into account in the course of a WLAN design.
Radio Frequencies, Transmit Power, and Antennae
A design consideration with respect to WLAN devices and antennae entails regulatory compliance with the permissible radio frequencies in a deployment region. For example, not all countries permit the use of the 5-GHz spectrum, which disqualifies 802.11a products from being deployed. However, with the 802.11g standard, which allows data rates up to 54 Mbps in the 2.4-GHz bands, a designer has choices.
Another regulatory element to consider is the Effective Isotropic Radiated Power (EIRP), which is derived from the combination of the following:
A wireless device's transmit signal power
Antenna gain
Antenna cable signal loss
NOTE
The European maximum values for EIRP differ from those in the United States.
These three elements are expressed in decibels (dB). A decibel represents 10 times the base 10 logarithm of the ratio between the measured power level and a reference power level, as expressed in the following formula:
10 × log10 (measured level/reference level)
If the measured level is 100 and the reference level is 1, then the base 10 log of that ratio is 2. When expressed in decibels, it is 20 dB (10 × 2). In practice, the dB symbol is combined with another symbol to represent what is being measured or what physical units are being used with a given product. When a device's transmit power is measured in milliwatts, the symbol dBm (m for milliwatts) is used to express power levels in decibels.
An easy rule to remember regarding decibels is that an increase of 3 dB represents a twofold increase in the power level, whereas a decrease of 3 dB corresponds to a halving of the power level. Assume that a device can transmit at power levels from 0 to 1000 milliwatts. Translated into decibels, the power levels for that device can be represented as follows:
1 milliwatt (mW) = 0 dBm
2mW = 3 dBm
4mW = 6 dBm
8mW = 9 dBm
32mW = 15 dBm
The maximum power of 1000mW is 30 dBm.
Through the use of the log function, decibels allow the representation of larger numbers in smaller increments. Power level configuration parameters for wireless devices are often expressed in decibels rather than in milliwatts or watts.
Antennae are critical elements of the wireless infrastructure because their characteristics determine the coverage range and performance of a wireless installation. For some WLAN products, you do not have a choice regarding an antenna because it comes integrated with a device. However, understanding antenna characteristics can help with the design of the WLAN topology. Consider next the antenna gain in the context of calculating EIRP. In addition to the gain, other key characteristics of antennae are the frequencies at which they operate, their direction, and their polarization.
EIRP and Antenna Gain
In contrast to transmit power levels, antenna gain is expressed in two ways:
dBi— In reference to an isotropic antenna. An isotropic antenna is the purely mathematical concept of a nondirectional antenna that radiates equally in all directions. When compared to itself, it has a 0-dB power loss/gain, making it easy to be used as a reference. The simple rule regarding antennae is that the greater the dBi, the higher the antenna gain, which can translate into a greater and more accurate coverage area for a WLAN. An antenna gain of 20 dBi represents 100 times more signal strength than the ideal isotropic reference. A gain of 30 dBi represents 1000 times more than the reference.
dBd— In reference to a dipole antenna. Dipole antennae are practically possible, and their gain is expressed in decibels as dBd (the second “d” is for dipole). A dipole antenna also has a 2.14-dB gain (which is often rounded up to 2.2) over an isotropic antenna. To convert a dBd antenna rating to a dBi rating, add 2.14 (or 2.2) to the dBd rating.
To calculate the EIRP, consider a hypothetical device with a maximum of 24 dBm (256mW maximum power), an antenna with a 28-dBi gain, and a cable from the device to the antenna with a 5-dB loss. EIRP is the sum of the decibels for all of those components, or 47 dB (24 + 28 ∠ 5). To derive the final EIRP value, loss means subtraction rather than addition because loss represents a negative value.
Always review the maximum permissible EIRP levels in a deployment region before proceeding with a more complex or large-scale WLAN implementation.
Antenna Radio Frequencies
For WLAN applications, the radio frequencies are either in the 2.4- or 5-GHz bands, which are broken into multiple channels. The 802.11b standard defines in the 2.4-GHz range 14 direct sequence (DS) channels that are 22 MHz wide, with 11 of these channels approved for use by the Federal Communications Commission (FCC) in the United States. In other countries or regulatory domains, the approved channel use might coincide or vary from that in the United States.
Due to the 5-MHz separation between the center of frequencies in the 802.11b channels, they overlap and interfere with each other if adjacent channels are used at an installation. Four successive channels are required between any two nonoverlapping channels, thus allowing for three nonoverlapping channels (1, 6, and 11) out of the 11 approved by the FCC.
WARNING
Use of overlapping channels in adjacent AP installations can lead to performance degradation and is not recommended. As a general rule, the 2.4-GHz antennae support all of the 802.11b-defined channels, but that's not necessarily the case with the 5-GHz antennae.
The 802.11a standard defines 12 nonoverlapping channels in the three Unlicensed National Information Infrastructure (UNII) 5.4-GHz bands for use in the United States. From the regulatory perspective, design considerations that apply to the 2.4-GHz frequencies are similar to those that apply to the 5-GHz bands.
NOTE
As a WLAN designer, you should always remember the frequency spectrum regulatory issues when planning multinational wireless deployments. Always check the spectrum regulatory requirements at the regulatory domain or the country of WLAN deployment.
Antenna Polarization
Antenna polarization refers to the orientation of the electric field component (with respect to the Earth's surface) of the transmitted electromagnetic waves (horizontal or vertical) and whether the electromagnetic radiation remains in a single plane (linear) or multiple planes (circular or elliptic). Most WLAN equipment antennae are linear and vertically polarized, with some exceptions that offer either horizontal or vertical polarization for the same model. Horizontally polarized antennae are used in broadcast TV in the United States.
The critical design issue regarding polarization is that the communicating antennae should have the same polarization. Although in practice there might be some performance-impacting exceptions to this guideline, don't make the mistake of using antennae with different polarization for longer line-of-sight (LOS) point-to-point communications between wireless bridges.
Antenna Direction
An antenna's type and style determine the direction of the radiation of the radio waves. Table 6-1 explores the characteristics of the two antenna categories—omnidirectional and directional—with respect to their direction of radiation.
| Antenna Direction Category | Style | Action | Deployment Scenarios |
|---|---|---|---|
| Omnidirectional | Typically shaped as a circular rod. The standard dipole antenna (the classic rubber duck) is an example. | Radiates in all directions or in a 360-degree pattern. | “Base stations” or APs in an office building where clients want to connect from any direction.[*] |
| Directional | Yagi, patch, sector, and parabolic dish. | Radiation is focused in a specific direction | Longer outdoor point-to-point bridge links.[*] Should work for a short-distance indoor bridge link. When mounted in a ceiling corner, it might be a perfect choice for a group of stationary wireless clients residing within its pattern of radiation. |
[*] These are typical deployment scenarios. An actual deployment scenario is always a function of user requirements and might depend on the applications and location of the communicating devices.
Because an antenna does not amplify the received signal from a wireless device but rather retransmits it in a different form, a directional antenna of comparable gain as an omnidirectional one will have a longer range. That is a result of all of the radiated energy being confined to a smaller angle of radiation.
Additional Antenna Considerations
Certain WLAN deployment environments are prone to multipath distortion or interference that occurs when a radio signal takes multiple paths while traveling between a source and a destination (a client and an AP, for example).
Multiple paths for an RF signal can result in a portion of the signal taking a longer path and experiencing a longer delay than the rest of the signal. This situation can lead to bit errors that eventually result in packet retransmissions and degraded performance. Environments with a lot of metal structures are most prone to multipath distortion. Consider the use of diversity antennae in those kinds of environments, which should be identified for the potential of multipath distortion via an RF site survey.
Diversity antennae are intended to overcome multipath distortion (or at least to minimize it) by utilizing two antennae to cover the same area. The antennae should not be so far apart that they cover distinctly different areas or be so close together that diversity does not occur. The separation between antennae in a diversity scenario should be in the range of 1 to 8 feet or approximately 0.3 to 2.4 meters.
A wireless device, such as an AP equipped with a diversity antenna, automatically switches between the two antennae to establish the best possible connection to a client without any intervention on the part of a user. Support for diversity antennae in an AP might be a design consideration if the user needs robust coverage in areas prone to diversity interference.
Access Points and Wireless Bridges
From the network infrastructure perspective, a wireless AP is effectively equivalent to a LAN access switch that is aggregating a variety of wireless clients. LAN switch characteristics, as they relate to network design, are considered primarily at the data link layer and are addressed in Chapter 3, “Network Infrastructure Requirements for Effective Solutions Implementation.” But similar to a switch, an AP also functions at the physical layer. In the case of an AP, the antenna characteristics, as presented in the preceding section, must be considered instead of cabling-related issues.
Another big difference between a typical AP and a LAN switch is that an AP is going to have a single 10/100 Mbps uplink to a LAN switch. That link must be shared by all of the wireless clients attached to an AP if they are accessing resources on the wire-based LAN. At first glance, the uplink capacity might seem to be a performance bottleneck. But consider that 802.11a and 802.11g specify maximum data rates of 54 Mbps, with 802.11b being lower. With the availability of dedicated Gigabit Ethernet to the desktop, a 54-Mbps data rate for a single client (or a group of clients sharing the total AP bandwidth) and a 10/100 uplink might seem archaic and slow. However, keep in mind that 10 Mbps Ethernet was the norm nearly a decade ago, and even 10 Mbps dedicated bandwidth is a screamer for many SMB applications.
Typically, servers might not be considered candidates for connecting to an AP, but nothing prevents a wireless notebook computer or a desktop from being configured as a server, even for temporary use. Naturally, large file transfers going on all day long, heavy-duty video editing, or graphic design work demanding high bandwidth would not be the ideal applications to be deployed over a wireless LAN with a limited bandwidth capacity. But as I've stated many times before, well-defined user requirements drive deployment.
As a function of the user requirements, an AP might have to support VLANs, VLAN trunking over the uplink, mobile IP if multiple subnets are in use in a larger WLAN deployment, quality of service (QoS) for prioritizing voice or other time-sensitive applications, and a variety of security protocols to ensure the security of the communications. APs must of necessity offer good support for authentication and data encryption. And if there is one place where you should be looking at security features, it is on APs.
In the early days of networking, bridges were either local or remote. Remote bridges operated in pairs to create a point-to-point link extending the LAN into a WAN or a metropolitan-area network (MAN), whereas a single local bridge could be deployed to improve network performance through segmentation of shared LAN segments. Wireless bridges also come in two flavors. Wireless bridges without additional attributes are a form of what used to be (and in some SMB installations still are) remote bridges. Wireless workgroup bridges, on the other hand, allow a small wire-based workgroup to connect to the workgroup bridge, which in turn can form an association with an AP. Figure 6-1 illustrates the two types of wireless bridges.
Unlike APs, bridges do not directly aggregate wireless clients. Instead, they extend a wireless or a wire-based LAN topology as a function of their physical layer characteristics and user requirements.
The physical layer characteristics that determine the ability of bridges to extend the size of a LAN include the following:
Radio frequency at which they operate
Transmit power settings
Antenna gain, direction, and style
In addition, bridges might need to support the following features, which are normally supported on the nonworkgroup bridges:
Wireless Routers
Think for a moment about the fundamental appeal of a wireless router—integration that is aimed at a specific market segment. A router that combines a WAN interface (cable/digital subscriber line [DSL]) and a standards-compatible AP to aggregate wireless clients has its place in networking, but most likely at the SOHO or SMB workgroup level.
Given the significant performance differences between maximum throughput capabilities of a wire-based versus a wireless link, a core switch/router that is installed in a highly secure data center and supporting Gigabit Ethernet and/or OC-192 interfaces is not likely to be incorporating an AP for aggregating wireless clients. You never know, though, what the future might bring! On the other hand, an access router incorporating a small switch (for a server or another wire-based device), an AP, a WAN interface for Internet access, and perhaps even a firewall might be a perfect wireless solution for a small branch office, an SMB workgroup, or a SOHO-type business.
Several companies offer wireless routers aimed at the home networking and SOHO markets. In mid-2003, Cisco Systems acquired the privately held Linksys Group, Inc., which specializes in home networking equipment. Wireless products from Linksys include several router models, client NICs, and even a wireless camera.
Wireless Clients
Wireless clients include notebook computers with built-in wireless adapters, Peripheral Component Interconnect (PCI) NICs that can be installed in servers or desktop workstations with a PCI slot, PC cards that nicely complement a notebook with PC slots but without a built-in wireless adapter, handheld personal digital assistants (PDAs), phones, cameras, universal serial bus (USB)-based wireless clients for ad hoc or AP-based communications, and more.
Given the availability of varying WLAN standards, the number of security protocols, and the number of vendors in the wireless networking field, the biggest challenge that you will probably face when considering wireless clients' deployment will have to do with compatibility and interoperability between the WLAN infrastructure and the potential clients.
The issues of compatibility and interoperability come into play even more when certain proprietary techniques are implemented in the wireless products to improve their performance or security. There is, however, no cause for alarm because the networking field, in general, is full of gear from numerous vendors. It is almost impossible to find an SMB installation without some kind of a multivendor network solution. The key is to be aware of the protocol implementations in the clients and the infrastructure devices (APs, routers) with which they will associate.
If compatibility problems persist for the same protocol implementations from different vendors, consider a single-vendor solution or be prepared for some lengthy discussions with technical support personnel of multiple vendors.