Wireless Network System Components

Wireless Network System Components

A wireless network consists of several components that support communications using radio or light waves propagating through an air medium. Some of these elements overlap with those of wired networks, but special consideration is necessary for all of these components when deploying a wireless network. Figure 2-1 illustrates these primary components.

A user can be anything that directly utilizes the wireless network. One of the most common types of user is a person. For example, a business traveler accessing the Internet from a public wireless LAN at an airport is a user. In some cases, however, the user might not be human. A robot, for example, might receive instructions over a wireless network from a central computer that controls a manu-facturing process. Because the wireless network exists to serve the user, the user is the component that receives the benefits of a wireless network. As a result, users are an important part of the wireless network.

The user initiates and terminates use of a wireless network, making the term end-user appropriate. Typically, a user operates a computer device, which often performs a variety of application-specific functions in addition to offering an interface to the wireless network.

Other users might actually be stationary, which means that they operate from one place for an indefinite period of time. An example of this type of user is someone working from a wireless computer in an office. The biggest difference between a stationary and portable user is that the stationary user will not require any form of roaming functions. Roaming functions are difficult to implement in some situations.

Many types of computer devices, sometimes referred to as clients, operate on a wireless network. Some computer devices might be specifically designed for users, whereas some computer devices are end systems. In generally, any computer device might communicate with any other computer device on the same wireless network. Figure 2-2 illustrates an assortment of computer devices for wireless networks.

Computer devices within a wireless network also include end systems such as servers, databases, and websites. For example, the http://www.cnn.com website includes news that someone can view from a public wireless LAN connection from a hotel room. Similarly, a clerk can wirelessly interface with a warehouse management system, which acts as an end-system computer device.

Users can adapt many existing computer devices to operate on a wireless network. A user, for example, can purchase and install a wireless network interface card (NIC) within his laptop to enable operation on a particular type of wireless network. Some devices, such as a wireless bar code scanner, operate only on a wireless network.

A computer device also has an operating system, such as Windows XP, LINUX, or MAC OS. The operating system runs software needed to realize the wireless network application. In some cases, the operating system has built-in features that enhance wireless networks. For example, Windows XP has the ability to automatically identify and associate with wireless LANs.

The network interface card provides the interface between the computer device and the wireless network infrastructure. The NIC fits inside the computer device, but external network adaptors are available that plug in and remain outside the computer device. Figure 2-3 shows examples of several types of wireless NICs.

Wireless NICs also comply with a specific form factor, which defines the physical and electrical bus interface that enables the card to communicate with the computer device. Again, the user must consider this to ensure that the chosen wireless NIC will fit within their computer device. The following is a summary of the different internal form factors available for wireless networks:

• Industry-Standard Architecture (ISA)— ISA has been around since the early 1980s. Because of this, the proliferation of the ISA bus has been significant. Despite its limited performance, nearly all PCs manufactured up until recently had at least one ISA bus. The ISA bus has failed, however, to advance at the pace of the rest of the computer world, and other higher-speed alternatives are now available. ISA doesn't impose too much of a performance impact on 802.11b wireless LANs. It's not advisable, however, to purchase new ISA cards because of the possibility of them becoming obsolete.

• Peripheral Component Interconnect (PCI)— The PCI bus is the most popular interface for PCs today and boasts high performance. Intel originally developed and released PCI in 1993, and it satisfies the needs of the recent generations of PCs for multimedia and graphics. PCI cards were the first to popularize "plug-and-play" technology, which makes it easy to install the NIC. PCI circuitry can recognize compatible PCI cards and work with the computer's operating system to set the configurations for each card. This saves time and prevents installation headaches for nontechnical users.

• PC Card— The PC Card was developed in the early 1990s by the Personal Computer Memory Card International Association (PCMCIA). The PC Card is a credit-card-sized device that provides extended memory, modems, connectivity to external devices, as well as wireless LAN capabilities to small computer devices such as laptops and PDAs. In fact, they are the most widely available NICs available. They are more popular than ISA or PCI cards because of use in a growing number of laptops and PDAs.

  It's possible to share a PC Card with a desktop PC by using an adaptor that converts a PC Card into a PCI card. This allows purchasing one NIC for use in both types of computers. You can take the PC Card with you on a business trip— or home from work— and utilize the same card when back in the office using a PC. Some PDAs require a sled device that accommodates the PC Card and mounts underneath the PDA. This is the only way to add wireless network capability to some older PDAs. The combination of the sled, PC Card and PDA, however, adds a lot of bulk and weight that depletes the usability.

• Mini-PCI— A Mini-PCI card is a smaller version of a standard desktop PCI card and fits well within small, mobile computer devices. It has all the same features and functionality of a normal PCI card, but is about one quarter the size. Mini-PCI cards are integrated within laptops as an option to buyers. A strong advantage of this form of radio NIC is that it frees up the PC Card slot for other devices, such as memory extenders and graphics accelerators. In addition, manufacturers can provide Mini-PCI–based wireless NICs at lower costs. The Mini-PCI card is not without disadvantages, however. The replacement of a Mini-PCI card typically requires the disassembly of the laptop, which might void the manufacturer's warranty. Mini-PCI cards might also lead to lower performance because they require the computer to do some, if not all, of the processing. Despite these drawbacks, the Mini-PCI card is becoming a solid technology in the wireless laptop world.

• CompactFlash— SanDisk Corporation first introduced CompactFlash (CF) in 1994, but wireless NICs were not available in CF form factors until recently. A CF card is small, weighing half an ounce, and is less than half the thickness of a PC Card. It also holds only one quarter the volume of PC Card radio card. The CF cards draw little power, which enables the batteries to last longer than devices using PC Cards. Some PDAs come with direct CF interfaces, which results in a lightweight and compact wireless PDA. If the computer device doesn't have a CF slot, you can purchase an adapter so that the CF card will fit into a standard PC Card slot. A CF radio card is definitely the way to go, especially for compact computing devices.

As Chapter 3, "Radio Frequency and Light Signal Fundamentals: The Invisible Medium," discusses in detail, a wireless NIC includes an antenna that converts electrical signals to radio or light waves for propagation through the air medium. Antennae employ many structures, and they can be external, internal, permanent, or detachable. The antenna for a PC Card, for example, generally attaches to the end of the card and protrudes out the side of the laptop.

Mini-PCI cards, however, might have an antenna that resides inside the outer edge of a laptop monitor. Some NICs have antennaes that are permanent, which have one particular propagation pattern. Other NICs allow the replacement of the antenna, which increases flexibility in choosing an antenna that best satisfies requirements.

Air serves many purposes, such as providing a basis for speech, enabling air travel, and sustaining life. Air also provides a medium for the propagation of wireless communications signals, which is the heart of wireless networking. Air is the conduit by which information flows between computer devices and the wireless infrastructure. Think of communication through a wireless network as similar to talking to someone. As you move farther apart, it's more difficult to hear each other, especially when a loud noise is present.

With wireless networks, the air medium supports the propagation of radio and light waves that travel from one point to another. These types of signals have been in use for more than 100 years, but they are still somewhat mysterious and not well understood by most computer professionals. Chapter 3 provides details on signal characteristics and impairments that relate to the air medium.

The infrastructure of a wireless network interconnects wireless users and end systems. The infrastructure might consist of base stations, access controllers, application connectivity software, and a distribution system. These components enhance wireless communications and fulfill important functions necessary for specific applications.

The base station is a common infrastructure component that interfaces the wireless communications signals traveling through the air medium to a wired network—often referred to as a distribution system. Therefore, a base station enables users to access a wide range of network services, such as web browsing, e-mail access, and database applications. A base station often contains a wireless NIC that implements the same technology in operation by the user's wireless NIC.

Base stations go by different names, depending on their purpose. An access point, for instance, represents a generic base station for a wireless LAN. A collection of access points within a wireless LAN, for example, supports roaming throughout a facility. The NIC within a user's computer device connects with the nearest access point, which provides an interface with systems within the infrastructure and users associated with other access points. As the user moves to a part of the facility that's closer to another access point, the NIC automatically reconnects with the closest access point to maintain reliable communications.

Residential gateways and routers are more advanced forms of base stations that enable additional network functions. The gateway might have functions, such as access control and application connectivity, that better serve distributed, public networks. On the other hand, a router would enable operation of multiple computers on a single broadband connection.

As show in Figure 2-4, a base station might support point-to-point or point-to-multipoint communications. Point-to-point systems enable communications signals to flow from one particular base station or computer device directly to another one. This is a common infrastructure for supporting long-range wireless communications links. For example, a wireless Internet service provider (WISP) can use this system to transport communications signals from a base station at a remote site— such as a home or office— to a base station near a communications facility.

In the absence of adequate security, quality of service (QoS), and roaming mechanisms in wireless network standards, companies offer access-control solutions to strengthen wireless systems. The key component to these solutions is an access controller, which is typically hardware that resides on the wired portion of the network between the access points and the protected side of the network. Access controllers provide centralized intelligence behind the access points to regulate traffic between the open wireless network and important resources. In some cases, the access point contains the access control function.

Access controllers apply to a wide range of applications. In a public wireless LAN, for example, an access controller regulates access to the Internet by authenticating and authorizing users based on a subscription plan. Similarly, a corporation can implement an access controller to help a hacker sitting in the company's parking lot from getting entry to sensitive data and applications.

The users of access controllers realize the following benefits when deployed with thin access points:

• Lower Costs— Access points with limited functionality cost less, which generally results in lower overall system costs. This is especially true for networks requiring a larger number of access points, such as an enterprise system. The use of thin access points results in cost savings of approximately $400 per access point. In larger networks, this savings far outweighs the additional cost of an access controller, which costs $5000 on the average.

• Open Connectivity— Smart access points offer enhancements related to security and performance to the basic wireless connectivity that wireless network standards offer. The problem in many cases is that these enhancements are only possible if the user devices implement a wireless NIC made by the same vendor as the access point. This significantly reduces the openness of the system and limits the selection of vendors. On the other hand, thin access points can easily communicate using the basic wireless network protocol with wireless NICs made by multiple vendors, while the access controller transparently provides enhancements.

• Centralized Support— An advantage of placing the smarts of the network in an access controller is that the system is easier to support, primarily because fewer touch points are in the network. If all of the intelligence of the network is within the access points, support personnel must interface with many points when configuring, monitoring, and troubleshooting the network. An access controller enables the access points to have fewer functions, reducing the need to interface with the access points when performing support tasks.

Access controllers often provide port-based access control, allowing administrators to configure access to specific applications on a per-user basis. The port, which is actually a number (such as 80 for http), corresponds to a specific type of application. For example, an access controller can block access to port 80, which forces a user to log in before being able to browse web pages. After users enter their username and password, the access controller will validate their identity through an authentication server. The network application could, as an alternative, use digital certificates for authentication purposes. This function regulates the user access to the protected network.

• Authentication— Most access controllers have a built-in database for authenticating users; however, some offer external interfaces to authentication servers such as Remote Authentication Dial-In User Service (RADIUS) and Lightweight Directory Access Protocol (LDAP). For smaller, private networks, an internal database might suffice. For enterprise solutions, however, external and centralized authentication servers provide better results.

• Encryption— Some access controllers provide encryption of data from the client to the server and back, using such common methods such as IPSec. This provides added protection beyond what the native wireless network standard provides. Some of these features, however, are also part of web browsers.

• Subnet Roaming— In order to support roaming from one network to another, access controllers provide roaming across subnets without needing to re-authenticate with the system. As a result, users can continue utilizing their network applications without interruption as they roam about a facility. This feature is especially useful for larger installations where access to the network for specific users will span multiple subnets.

• Bandwidth Management— Because users share bandwidth in a wireless network, it's important to have a mechanism to ensure specific users don't hog the bandwidth. Access controllers provide this form of bandwidth management through the assignment of user profiles based on required QoS levels. A profile specifies the types of services, such as web browsing, e-mail, and video streaming, as well as performance limits. For example, an unsubscribed visitor attempting to utilize a public wireless LAN could classify as fitting a "visitor" profile, which might only allow access to information related to the local hotspot. A subscriber, however, could have a different role that allows him to have a broadband Internet connection.

Web surfing and e-mail generally perform well over wireless networks. All it takes is a browser and e-mail software on the client device. Users might lose a wireless connection from time to time, but the protocols in use for these relatively simple applications are resilient under most conditions.

Beyond these simple applications, however, special application connectivity software is necessary as an interface between a user's computer device and the end system hosting the application's software or database. Applications could be warehouse management software running on an IBM AS/400, a modeling application located on a UNIX box, or a time-management system residing on an old mainframe system. The databases are part of a client/server system where part, or all of the application software, resides on the client device and interfaces with a database such as Oracle or Sybase. In these cases, application connectivity software is important in addition to access points and controllers to enable communications between the user's computer device and the application software or databases located on a centralized server.

• Terminal Emulation— Terminal emulation software runs on a computer device, making the device operate as a terminal that provides a relatively simple user interface to application software running on another computer. The terminal merely presents screens to the user and accepts input rendered by the applications software. For example, VT220 terminal emulation communicates with applications running on a UNIX host, 5250 terminal emulation works with IBM AS/400-based systems, and 3270 terminal emulation interfaces with IBM mainframes.

  The advantage of using terminal emulation is its low initial cost and changes made to the application automatically take affect when the user logs in. Wireless systems using terminal emulation, however, might not be able to maintain continuous connections with legacy applications, which have timeouts set for more reliable wired networks. Timeouts will automatically disconnect a session if they don't sense activity within a given time period. As a result, IT groups might spend a lot of time responding to end-user complaints of dropped connections and incomplete data transactions. Therefore, implementing terminal emulation can have a disastrous effect on long-term support costs.

• Direct Database Connectivity— Direct database connectivity, sometimes referred to as client/server, encompasses application software running on the user's computer device. With this configuration, the software on the end-user device provides all application functionality and generally interfaces to a database located on a central server. This enables flexibility when developing applications because the programmer has complete control over what functions are implemented—and is not constrained by a legacy application located on a central computer. Direct database connections are often the best approach when needing flexibility in writing the application software. A problem, however, is that the direct database approach depends on the use of Transmission Control Protocol/Internet Protocol (TCP/IP), which is not well-suited for communications across a wireless network.

• Wireless Middleware— Wireless middleware software provides intermediate communications between user computer devices and the application software or database located on a server. (See Figure 2-5.) The middleware—which runs on a dedicated computer (middleware gateway) attached to the wired network—processes the packets that pass between the user computer devices and the servers. The middleware software primarily offers efficient and reliable communications over the wireless network while maintaining appropriate connections to application software and databases on the server through the more reliable wired network. Sometimes this is referred to as session persistence.

  Figure 2-5. Wireless Middleware Efficiently Interconnects Computer Device Applications to Hosts and Servers

  

  
  

- Optimization techniques— Many middleware products include data compression to help reduce the number of packets sent over the wireless link. Some implementations of middleware use proprietary communications protocols, which have little overhead as compared to traditional protocols, such as TCP/IP.

- Intelligent restarts— With wireless networks, a transmission can be unexpectedly cut at midstream. Intelligent restart is a recovery mechanism that detects the premature end of a transmission. When the connection is reestablished, the middleware resumes transmission from the break point instead of at the beginning. This avoids errors from occurring in applications that utilize databases.

- Data bundling— Some middleware is capable of combining smaller data packets into a single large packet for transmission over the wireless network, which can help lower transmission service costs of WANs. Since some wireless data services charge users by the packet, data bundling results in a lower aggregate cost.

- Screen scraping and reshaping— The development environment of some middleware products allows developers to use visual tools to shape and reshape portions of existing application screens to more effectively fit data on the smaller display of some non-PC wireless devices, such as PDAs and bar code scanners.

- End system support— Wireless middleware interfaces with a variety of end system applications and databases. If clients need access to tomultiple types of applications and databases, wireless middleware acts as a concentrator. For example, a user can use the middleware connection to interface with applications on an AS/400 and UNIX box simultaneously without needing to be concerned about running the correct terminal emulation software.

A wireless network is seldom entirely free of wires. The distribution system, which often includes wiring, is generally necessary to tie together the access points, access controllers, and servers. In most cases, the common Ethernet comprises the distribution system.

The IEEE 802.3 standard is the basis for Ethernet and specifies the use of the carrier sense multiple access (CSMA) protocol to provide access to a shared medium, such as twisted-pair wiring, coaxial cable, and optical fiber. CSMA is the predominant medium access standard in use today by both wired and wireless networks.

CSMA enables sharing of a common medium by allowing only one NIC to transmit information at any particular time. This is similar to a meeting environment where people (like NICs) speak only when no one else is talking. This gives each person responsibility in a way that distributes speaking decisions to each person. If more than one person talks at the same time, a collision occurs, and each person needs to take turns repeating what he said.

Ethernet employs twisted-pair wiring, coaxial cable, and optical fiber for interconnecting network devices, such as access points and other distribution equipment. The use of coaxial cables in older wired LANs was common 10 years ago, but today most companies use twisted-pair wiring and optical fiber. The Electronic Industries Association (EIA) and Telecommunications Industry Association (TIA) specifies Category 5 (referred to as Cat 5) twisted-pair wiring, the most popular of all twisted-pair cables in use today with Ethernet.

There are also other variations of twisted-pair wiring. Enhanced Cat 5 (referred to as Cat5e) makes use of all four pairs of wires to support short-range Gigabit Ethernet (1000 Mbps) connectivity. It is also backward compatible with regular Cat 5. Cat 6 and Cat 7 cable are now available, bringing more bandwidth and range to copper-based Gigabit Ethernet networks. Cat 7 cable features individually shielded twisted pairs (STP) of wires, making it ideal for installation in locations where there is a high potential for electromagnetic interference.

The following are specific types of twisted-pair options for Ethernet common to wireless LAN distribution systems:

• 10BASE-T— 10BASE-T is one of the 802.3 physical layers and specifies data rates of 10 Mbps. A typical 10BASE-T cable uses two of the four pairs within a Cat 5 cable for sending and receiving data. Each end of the cable includes RJ-45 connectors that are a little larger than the common RJ-11 telephone connector used within North America.

  The advantage of having extra pairs of wires open is support for other uses, such as Power-over-Ethernet (PoE). This is a mechanism in which a module injects DC current into the Cat 5 cable, enabling you to supply power to the access point from the communications closet. PoE often eliminates the need for having an electrician install new electrical outlets at every access point. For larger networks, definitely consider the use of PoE.

• 100BASE-T— Another 802.3 physical layer, 100BASE-T supports data rates of 100 Mbps. Similar to 10BASE-T Ethernet, 100-Base-T uses twisted-pair wiring, with the following options:

  - 100BASE-TX uses two pairs of Cat 5 twisted-pair wires.

  - 100BASE-T4 uses four pairs of older, lower-quality (Cat 3) twisted-pair wires.

  Most implementations today use 100BASE-TX cabling. As with 10BASE-T, PoE can make use of unused pairs of wires. 100-Base-T4 was popular when needing to support 100-Mbps data rates over the older Cat 3 cabling, which was prominent during the early 1990s.

• Optical Fiber— Optical fiber is more expensive than twisted pair, but fiber can be cost effective because it supports gigabit speeds and has a range of up to two kilometers. Instead of using the traditional electrical-signal-over-copper-wire approach, optical fiber cable uses pulses of light over tiny strips of glass or plastic. This makes optical fiber cable resistant to electromagnetic interference, making it valuable in situations where electronic emissions are a concern. In addition, it's nearly impossible to passively monitor the transmission of data through optical fiber cable, making it more secure than twisted-pair wiring.

  With wireless LANs, optical fiber is a possible solution for reaching an access point located beyond a 100 m from a communications closet. This requires the use of an expensive pair of transceivers, however, which transforms electrical signals into light (and vice versa). One issue when dealing with optical fiber cable is the difficulties in splicing cables. You must work with glass or plastic materials that require precise alignment. You need special tools and training to make effective optical fiber cables. You should purchase precut fiber cables to avoid problems that are difficult to troubleshoot.

As with other types of networks, enterprise wireless networks require effective management that ensures user needs are met over the life of the network. A network management system, which involves both people and software tools, satisfies this need. The following are functions that management systems should provide.

The security element involves mechanisms that avoid the compromise or damage of network resources, such as databases and e-mail messages. This includes enforcing security policies for the configuration of the wireless network in a way that counters issues related to the propagation of wireless signals. For example, policies could specify the use of a particular type of encryption to ensure a mischievous person can't receive and decode e-mail messages being sent between a user and an access point.

For more details on wireless network security methods, refer to Chapter 8, "Wireless Network Security: Protecting Information Resources."

The help desk provides the first level of support for users. A user having difficulties with a wireless connection should know how to reach the help desk. Users often have problems with association or experience erratic performance.

Help desk personnel are capable of solving simple connection problems, such as assisting the user configure a radio card and operating system to comply with specific wireless network policies. The help desk should have a communications interface with more advanced support functions, such as maintenance and engineering, to solve more complex problems that arise from contact with users.

Configuration management consists of controlling changes made to the wireless network architecture and installed system. Changes might consist of installing or moving access points, changing access point parameters, and updating firmware. Because of the dynamic nature of wireless networks, changes are more common than with wired networks.

An enterprise should review all wireless network modification proposals that impact the performance or security of the network. This review enables a company to take into account relevant implications that involve additional costs and use of resources. The company should implement an independent design review process that evaluates each proposed wireless network solution and verifies conformance to a common architecture and support elements. The verification should, for example, include reviewing access point placement, radio frequency channel assignments, and security settings.

Network monitoring includes continuously measuring various attributes of the wireless network, including access point utilization and user traffic routes through the distribution system. This plays a key role in proactively managing the wireless network to support a growth of users and solve issues before they hamper the performance and security of the network.

In most cases, coverage from base stations overlaps, and users will likely associate with another base station at lower performance if the primary access point is not available. Network monitoring tools, however, will notice the outage immediately and alert the appropriate support person. If possible, a company should integrate the wireless network monitoring function with tools already in use in the existing corporate network. This simplifies operational support.

The reporting element offers information regarding various aspects of the wireless network, including usage statistics, security alarm notifications, and performance. These reports are necessary for managers to effectively gauge the operation of the network and make decisions on changes. This reporting should, at a minimum, indicate potential breaches of security, inoperable access points, and utilization. This type of information should be available to all operational support functions, such as the help desk, maintenance, and engineering.

The engineering element provides advanced technical support for reengineering the wireless network to include newer technologies and solve problems to ensure effective performance and security. Ordinarily, the company or group that designs the initial wireless network will perform the engineering functions. The engineering function should review and verify compliance of designs with the common architectural design. In addition, the engineering group should continually monitor the evolution of wireless network technologies and products to ensure effective migration in a manner that meets growing network utilization.

The maintenance element repairs and configures the wireless network, including replacing broken antennae, setting channels on access points, and re-evaluating radio wave propagation. Some maintenance tasks might result from the engineering support function. For example, engineers might find the need to install an additional access point in an area where new coverage is necessary. In this case, maintenance personnel would install the access point at a location that the engineer identifies.

The maintenance group should also periodically perform coverage tests to ensure that the access points are properly covering the facilities at applicable levels of performance. This is necessary as the company modifies the structure of the facilities, which changes the radio wave propagation characteristics. If discrepancies are found, the maintenance group should report findings to the engineering group for resolution.