Chapter 29

Adding Wireless Networking

What is wireless networking? Although at first you might think of wireless networks in terms of simple LANs and desktop machines, in reality a wireless network can consist of almost any kind of computing machine and can extend over any area. Essentially, a wireless network transmits data along a certain frequency through the air rather than through a cable. Wireless networks include

These are just examplesóafter reading this chapter, you'll probably start recognizing wireless networks in other places as well.

The size of the network and the transmission medium are irrelevant to the definition of a wireless network. The main thing that most wireless networks have in common is a need for mobility. Few wireless networks evolve from traditional wired LANsómore often, they stem from a unique need for networking components that are difficult or impossible to connect in any other way.

Much current wireless networking technology uses radio frequency (RF) signals to transmit data, but microwaves and infrared signals are also part of wireless technology. In this chapter, we'll talk about various kinds of wireless technologies and topologies that are most common among commercial systems, potential applications, and some policy and technical issues that affect the development of wireless networks.

Pros and Cons of Wireless Networking

Like networking in general, wireless networking has things to recommend it, as well as definite detriments. Whether the pros outweigh the cons depends on your particular needs.

Pros

Most of the positive things about wireless networking stem directly from the lack of cabling to mess with. No cables potentially means the following:

Networks in Inaccessible Areas

In parts of the world where there either is little cable infrastructure or it has been damaged by war or environmental disaster, wireless communications can keep remote sites in touch with each other. How do you connect a remote mining site to the city office to discuss equipment needs? Or coordinate rescue efforts after an earthquake? In these cases, a cable infrastructure either never existed or no longer works. The only way to expeditiously share information in such circumstances might be a wireless network.

Easier Network Setup

If a network's components (PCs and peripheral devices) are fluid, then a wireless configuration can make setting up the network much easier. This can work even if the network is a hybrid of wired and wireless connections. For example, if you have data on your notebook computer that you need to print out, a wired network provides you with a couple of options for getting that data from the notebook to the printer. You could sever the printer's usual connection and plug it into the notebookóor if the notebook had a docking station and NIC, you could plug the notebook into the network and thereby connect to the printer. If neither of those options were possible, you could copy the files to be printed to a PC that did have access to the printeróassuming that the application software used for the files was on the wired PC as well as the notebook. No matter which method you use, though, it's something of a hassle and requires a bit of setup time.

An easier solution would be an wireless connection between the two (assuming that both the notebook and printer were equipped for it). That way, you could send a job to the printer with no cable swapping, and without shoving components aside to make room. Just make sure that the printer and notebook are fairly close to each other with no obstructions and that the two components can be connected via a virtual cable.

This example illustrates the next point, as wellóprinter sharing without a network. If the printer is wireless-equipped (such as HP's 5P) and the notebook also is (such as IBM's ThinkPad), then one computer can connect to the printer via the parallel port and another via the infrared port. The PC connecting to the printer via the parallel port doesn't have to be wireless equipped at all, unless the two computers need to be able to share files. (We'll talk about hybrid networks, or networks with both wired and wireless components, later in this chapter).

Networking Mobile Clients

Two trends of the 1990sóincreasing worker mobility and increasing dependence on up-to-the-minute informationódrive mobile computing and make it vital for mobile computing clients to have network connections:

In all these situations, the clients need information that they can't get using a wired connection. (It's ridiculous to consider connecting fishing boats with cables.) If these mobile computing clients are to have access to networked information, their connection must be wireless.

Wireless Modem Connections

The advantages of a wireless modem connection are a little different from those of a wireless printer. With a wireless modem, you can potentially carry your office anywhere that you and your laptop goóa much more likely scenario than lugging a wireless printer around.

Wireless dialup capabilities offer great advantages for people who work a lot on the road and must rely on hotels to act as their office. For one thing, some hotels route all calls through a switchboard, so using a traditional modem isn't possibleóyou can't just unplug the telephone and plug your modem into the wall to get to an outside line. For another, printing on the road is often a major hassle and expense. Even if the hotel in which you're staying has a computer center (and many do not), printing charges in hotels range from fairly pricey to downright exorbitant. Via the modem, you can fax your document to the hotel's fax number and address it to yourself. The final product may be on thermal paper, but if all you need is hard copy and quality is not the main concern, then that's an easy and inexpensive way to get a printout.

Essentially, the advantages of wireless networking are mobility and flexibility. Wireless networking makes it possible to share information when the infrastructure or circumstances seem to make networking impossible. Wireless modems come either as PC-Card (PCMCIA) devices that fit into a slot on your PC, or as a separate device with an antenna, which can plug into any RS-232-compatible serial port.

***

If you're not sure that your serial port is RS-232 compatible, don't worry too muchómost are.

***

Cons

Mobility and flexibility aren't everything in a network. Wireless networks have some inherent disadvantages, although some are more perceived than actual. Generally speaking, the following hold true:

Even if some of these considerations (such as the security concerns) are more perceived than actual, perceived detriments often do as much damage to a product's acceptance as real ones, so they're worth considering.

Limits of Speed and Distance

Wireless communications are inherently limited by their lack of cabling. If the wireless medium is a radio-frequency signal, then it can travel for a long way, but it's slow (and often uncertain) compared to a wired connection. Slow, in this context, means that a high-speed wireless link is about 1.6 Mbps for a LAN and 64 Kbps for a WAN. Considering that most wired LANs run upward of 10 Mbps, and many WAN technologies (as discussed in chapter 4, "LANs versus WANs") start at more than 1 Mbps, wireless speeds don't compare very well.

If the medium is an infrared signal, then it's faster (as fast as a wired network), but it's limited in the distance that it can extend without regenerating the signal and is also dependent on line-of-sight between the device sending the signal and the one receiving it. Even trees can disrupt an infrared signal. We'll talk about why this is so a little later in this chapter, when we discuss how frequencies work.

Higher Cost

Wireless technology is relatively new and complex, so like other new computing technology it tends to be expensive. Wireless LAN network cards, for example, run about $400 for an Ethernet connection, while wired ones are available for $50 to around $100. Connection time for dialup wireless connections costs, too (but then so do some wired dialup connections). (This applies to wireless solutions that let you dial into your company's network.) Of course, wireless networks don't carry the cost of pulling cable, which is not inconsiderable (the exact amounts depend on your installation, requirements, and the part of the country you're in), and gets quite expensive for complex installations.

Environmental Constraints

Because of their physical structure, wireless networks are vulnerable to environmental factors that under normal circumstances don't affect wired networks, such as

Of course, not all these circumstances affect all wireless communications the same way, and newer technology has improved the situation. Environmental factors, however, are still a consideration.

For example, properly shielded wired connections aren't sensitive to other wired connections in the same area. You can have a jumble of cables in the same space, and as long as you're using properly shielded cables the signals won't interfere with each other under normal circumstances. (If your cable installation uses UTP, don't worryóthe twists in the wires contained in the cable help protect against interference.)

Wireless networks are different. Because the medium of transmission is a naked electronic signal using a certain frequency, two transmissions on the same frequencies in the same area will interfere with each other unless some sort of spectrum-sharing technology is used. This is really no different from the interference problem that you encounter with poorly shielded cables, such as the ribbon cables used to connect printers to PCs.

One of the challenges facing wireless networking is that the higher the frequency used (and thus the bandwidth), the more vulnerable the signal is to interference. There's a direct tradeoff between carrying capacity and stamina, which is discussed more in the following section.

Bandwidth Constraints

Increased use of higher frequencies and digital "messaging" techniques that permit greater throughput on the same bandwidth mean that there's no reason for wireless transmissions to be slower than wired ones, but current cost and policy issuesócombined with the fragility of the signal at higher frequenciesómean that for all practical purposes they are slower. Although you can get equivalent bandwidths from wired and wireless networks, it costs you far more to get the wireless speed because of the following:

How Secure Are Wireless Transmissions?

Anyone who's ever accidentally overheard a neighbor's conversation on their cordless telephone is aware that data traveling through the ether is pretty open to tapping. Anyone who is equipped with the proper equipment can intercept the signal. However, this is not a problem unique to wireless networksóit's just easier to intercept a wireless signal accidentally.

Before you decide that wireless communications (especially data) are more vulnerable to tapping than wired ones, consider the following points:

Health Concerns

The possibility of health risks posed by wireless transmissions are not yet fully known, but some dangers definitely exist. Unshielded microwave transmissions such as those used in some military communications can be dangerous (the reason that microwaves in public places don't carry warning signs anymore has more to do with improved shielding than with safer microwave signals). Therefore, it's possible that some people may be affected by RF signals since microwaves are just another part of the electromagnetic spectrum. For the moment, however, this doesn't seem to be a major problem for anyone except those who look directly into an infrared transmitter's LED.

A Brief Introduction to Wireless Technology

Before jumping into a discussion of wireless technology, it's helpful to understand the basic structure of this technology. In this section, we'll discuss the basics of what a frequency is, what affects its speed, and what the differences are between the various frequencies.

The basic thing to understand about wireless signals is that every signal operates at a certain frequencyóthat is, the number of oscillations per time unit that the signal makes. More oscillations means higher frequency, as you can see in figure 29.1.

Figure 29.1 A low frequency has fewer oscillations per minute than a high frequency.

The number of cycles per second is measured in hertz (abbreviated Hz). Most measurements we care about when discussing wireless communications involve megahertz (MHz) or gigahertz (GHz). A signal with a frequency of 30MHz, for instance, oscillates 30 million times per second. A higher hertz rate means that the signal has a greater bandwidth because more information (1s and 0s, represented by the waves) is included in the signal each second.

Frequencies are divided into discrete groups, called bands, based on their hertz rateóthus, based on their carrying capacity. Table 29.1 shows some of these bands.

Table 29.1 Important Radio Frequency Wireless Bands

The radio-frequency bands currently used most often are HF, VHF, and UHF, with the higher-frequency bands coming into greater use as the lower-capacity bands are either saturated or unable to carry the data loads that newer wireless transmissions demand. For instance, a frequency that can easily handle textual dispatching information might have problems carrying real-time video. Infrared (IR) transmissions use frequencies high above the levels listed in table 29.1óat the high end of the visible spectrum. Because of their higher frequency, IR transmissions have greater carrying capacity than RF transmissions but have a much shorter range. Generally speaking, the higher the frequency, the greater the bandwidth and the shorter the distance it will travel.

Because of their limited bandwidth, bands below HF, low frequency (LF) and medium frequency (MF), are not suited to wireless data transmissions.

Band Frequency Characteristics

How does a signal's frequency impact its ability to transmit data? We've already discussed one effect of higher frequency: more bandwidth. Another effect, however, is increased vulnerability to interference and attenuation. This is a direct result of the shorter wavelengths that you saw in figure 29.1. Shorter wavelengths make the signal less flexible, and therefore less able to "bend" around blockages and more prone to interference from other signals. This is just like the difference between a loosely-twisted rope and a tightly-wound rope. You can bend the looser rope around obstacles without much problem, but the tighter one won't bend as easily.

Although interference and attenuation have similar effects on data transmissionóthey distort itóthey're not the same thing. Interference is a stray electronic signal that distorts the signal being transmitted, possibly corrupting the data that it's carrying. Attentuation is the increasing weakness of an signal as it travels a longer distance than it's meant to; the fading signal can distort the data. Interference affects data transmission in the same way that too many people talking at once can make it difficult for you to understand one person. Attenuation adversely affects transmission in the same way as it's hard to understand a person across a room and speaking at face-to-face volume.

As you can see in figure 29.2, increased bandwidth means less signal staying power.

Figure 29.2 The greater a frequency's bandwith, the more quickly its signal degrades.

This is not an exact charting, as the effects of distortion on a signal are impossible to measure exactlyóeven if you deliberately distorted a signal and measured the result, how would you know that nothing else had affected the signal?óbut should give you the general idea.

Wireless transmission signals occupy part of a band, not the whole band. If a signaling technology uses frequencies in the VHF band, it is limited to only a few of those frequencies, not the entire run from 30MHz through 300MHz.

How Frequencies Transmit Data

That's the basic story of what frequencies are, but how does the data get from your PC to the frequency and from there to its destination? To begin with, the data is a series of electrical impulses at certain intervals, usually represented with 1s (impulse) and 0s (no impulse)óin this instance, let's say that it's an e-mail message. When you click the Send button (or equivalent thereof), you send the collection of 1s and 0s that make up your e-mail message and addressing information to a transmitter. Depending on what kind of transmitter it is, the transmitter converts the 1s and 0s to a related pattern of electrical signals that can be sent on a particular frequency band. For example, if your PC is equipped with a radio-frequency (RF) transmitter, then it will send the signals via either the HF or VHF bands shown in table 29.1. A reciever of the same type in the area can then pick up on the signal that you transmitted if it's tuned to the same frequency.

That's the easy part. The complex part comes when you try to limit the number of receivers (you'll note that, in this scenario, anyone who has a receiver tuned to the same frequency and is within range gets your message) or when you've got a number of devices that want to use the same bandwidth.

Conserving Bandwidth

You now know in general terms what frequency is, and how bands with different frequencies respond to environmental conditions and carry data. You might be wondering how, since there are a finite number of bands, wireless services can expect to expand. For that matter, how have they expanded as far as they have without using up all the available bandwidth?

This question is crucial to wireless networking. Like any other resource, radio frequencies are limited but the uses for them are not. When wireless transmissions were only for television signals and CB radios, this was less of a concern. (For the record, however, the CB craze in the 1970s caused a real bandwidth saturation problemóthe more crowded a frequency gets, the harder it gets to use it.) Today, though, wireless communications are increasingly important to many business applications. Where will the bandwidth needed to sustain growth in these services come from?

Isolating Frequencies

Part of the answer stems from the fact that RF signals don't extend forever. A particular frequency can be in use in more than one place as a time, as long as the usages are separated enough not to interfere with each other. For example, you can use the frequency 33MHz in Melbourne and Seattle at the same time, attenuation would prevent signals that originate at points so far apart from interfering with each other. By the time the Seattle signal gets to Australia, it is so weak as to be practically nonexistent.

The fact that you can use a frequency in more than one place at a time naturally has led to the creation of frequency "cells" that electrically isolate geographic areas from each other. These cells are hexagonal in shape and have a transmitter located at their center to handle transmissions. The closer a subscriber is to the transmitter, the stronger the signal is, but it won't cut off until the subscriber moves to a different cell. Within cell A, for example, a frequency can only be used by one signal at a time (that's for startersóthere's more to this story that we'll get to in a minute), but that frequency also exists in cell B, cell C, and so on. Electronic barriers between the cells keep the signals from extending beyond their cell. Adjoining cells can't use the same frequencies at the same timeóin cellular communications, they must be separated by at least six cellsóbut the cells allow some level of frequency reuse. The smaller the cells are, the greater the possible frequency reuse.

There is a tradeoff, however. Since it's obviously not acceptable to chop off signaling when a person making a cellular transmission crosses a cell's border, some kind of switching mechanism is in order. The switching mechanism is the problem. Switching the signal from cell A's transmitter to cell B's transmitter takes time and can result in a lost connection if something goes wrong. Thus, although smaller cells (like those used in personal communications services, discussed later in this chapter) mean greater frequency reuse, they also mean more switching.

Data Compression

To get more use out of available bandwidth, you can make sure that the data takes less time to travel to its destination, thus freeing the bandwidth sooner for other transmissions. Since the mid-80s, the algorithms for compression have dramatically improved due to better design and better processing hardware.

Channeling Methods

Another means of conserving bandwidth is to channel the frequencies so that more than one subscriber can use them at once. There are a number of different methods of doing this, but here we'll restrict the discussion to two that you're likely to encounter: time-division multiple access (TDMA) and code-division multiple access (CDMA). Although these channeling technologies function differently from each other, they have the same purpose: to let more than one signal use the same frequency in the same place at the same time.

TDMA uses the same multiplexing techniques that time-division multiplexing does. Bandwidth is split into a number of logical channels, and each signal gets one of the channels. For example, six devices can transmit data at 330MHz in the same place at the same time. The process actually involves extraordinarily rapid switching of the channel among its various "simultaneous" users so that each of them has the apparent sole use of the channel but is in fact sharing that channel with all its other users, occupying it for specific segments of time. Think of it as shuffling the users in and out of the channel at predefined intervals but doing the shuffling so quickly that it is transparent to them.

CDMA, also called spread-spectrum technology, works a little differently. Rather than dividing the frequency into perhaps six channels, as TDMA would, it mixes the six transmissions and sends them all as a heap. Each transmission has a unique digital code assigned to it that permits the recipient to sort out its own data from the other transmissions. CDMA works like a packet-switching network, the characteristics of which are discussed in chapter 4, "LANs versus WANs." This sounds confusing, but it's not that complicated. Think of a mailbox in an office building. If each office in the building has its own mailbox and must pay rent on it, that can get expensive and is unnecessary for those who don't constantly send mailómuch more efficient to have one big mailbox in the office building that everyone shares rent on. To keep everyone's mail from getting mixed up, each office color-codes its envelopes: Rutger's Insurance uses blue envelopes, Wilson Industries pink, and so forth. Thus, all of the offices can share transmission bandwidth (the mailbox) and jumble their signals together to get the most use out of the available bandwidth, but the data (that is, the mail) doesn't get confused because each piece is coded for the recipient).

The differences between the two modulation techniques are visible in figure 29.3.

Figure 29.3 TDMA allocates bandwith by time slot; CDMA allocates bandwith by packet.

Using cells, digital compression, some kind of logical bandwidth division, or a combination of the three, you can squeeze enough space out of existing bandwidth to hopefully fulfill the needs of wireless communications. If that doesn't work, there's always pressure from new services that require bandwidth insisting that "less crucial" users give up their bandwidth to others.

Frequency reallocation questions lead to highly-charged political wrangles. Bandwidth users with big audiences but less serious topics (like Saturday morning cartoons) must wrestle over frequencies with users who have more serious topics but much smaller audiences. Look for increasing numbers of slugfests as the competition for bandwidth increases.

Keeping Communications Secure

You'll recall that when you transmit a signal, any receiver in the area that's tuned into the same frequency can pick up the signal. This is great for radio stations but less desirable for your company's budget projection for next year. One method for getting around this is called frequency hopping. This is exactly what it sounds like: instead of using a single frequency for the duration of the transmission, the transmitter will first send an encoded key to the receiver, telling it what frequencies that it's going to use for the transmission. Then the transmitter will begin the transmission, jumping around in the spectrum to keep any receivers in the area from inadvertently picking up on the signal for more than a flash.

Ten years ago, frequency-hopping was largely limited to secure government and military wireless transmissions, and even today it's not required for all situations. Other security methods include encoding transmissions and password-protecting files. The high-end wireless carriers (like those providing microwave services) have government-approved measures that can keep your data secure.

RF Links

Historically, the majority of wireless technologies have relied upon RF signals to transmit data. There are several reasons for this. First, RF signaling was the earliest electronic form of wireless communications. Although original usage was restricted to lower frequencies, it was natural as needs changed to keep moving up to higher levels of RF signaling, rather than abandoning RF technology to explore new media. Related to this point, RF signaling was a medium with which most developers were familiar through radio and television, and presumably were more comfortable.

RF technology works for a number of different networking applications in both LANs and WANs. Inside, antennae in serial ports or transceivers in PC-Card slots can connect the portable computers to each other or, via a gateway machine, to a wired network. The speeds involved are not great (ranging from a couple of hundred Kbps to 1.6 Mbps). For transferring relatively simple files such as text-only messages, however, they're perfectly acceptable. Desktop machines also can connect to a wireless network, via an ISA (or PC-Card) wireless transceiver, or an external antenna.

Applications

Who might benefit from wireless networking? The next few sections introduce you to typical beneficiaries of wireless technology.

Companies That Maintain Inventory

Which would you rather do: perform inventory with a clipboard in your hand and then enter your tallies in the computer when you get back to your desk, or enter the tallies directly into the inventory database? The first method might work if you have a standard inventory form and don't have to try to decipher your abbreviations or do the math five times to make sure your numbers are right, but it's twice the work for the same result and takes longer to get an accurate count. You could take a stand-alone notebook computer to do inventory, but the updates would not be automatic. Sooner or later you'd have to copy the files from the notebook's hard disk to merge them with the main inventory database. Even having a network connection and docking station in the storage room might not work, because you can't carry the notebook with you when it's in a docking station.

If, however, your notebook had a wireless connection to the inventory database server, you could do inventory and update the database at the same time. You would just type in the entries as you went.

Companies with a Mobile On-Site Workforce

I bought a car not too long ago, and one of the major hassles about the purchase process was that every time the salesperson wanted to run some numbers through the auto dealership's database, he had to walk to a terminal at the other end of the room and do the work there. This was irritating not only because he kept disappearing for extended periods of time, but because a couple of times he forgot the numbers that he wanted to input and had to come back to get them. It was annoying for me, and probably frustrating for him as well. Since only a few salespeople had their own desks, it wasn't practical for them to have wired computers on the desktops, since on a busy day they'd be stuck waiting for a desk with a computer, or hovering unprofessionally over each other, waiting for a crack at a computer.

If each of the salespeople had a portable PC, on the other hand, they could access the information they need, and make their comparisons without having to leave the desk. This could expedite the entire process, and it certainly would make customers happieróif you must wait, it's easier to wait when you know what's going on, not just that the salesperson has disappeared again.

Companies with a Mobile Off-Site Workforce

Many salespeople spend most of their time either at the client's site or traveling to it. A wireless WAN that permits these salespeople access to the office network might make a sale. The principle is the same as that of the on-site salesperson: information available at your fingertips, so that the customer doesn't have to make another appointment to get readily-available facts, results in happier customers and a more likely sale.

For that matter, when those mobile salespeople come back on-site, wireless technology can make it easier for them to connect to the network and even print. A wired/wireless bridge or gateway machine can let a laptopóas long as it's equipped with a transceiver and client softwareóplug into the network without having to find a docking station.

Equipment Used in RF Wireless Networking

The components of a wireless network are not very different from those used in a wired network. At a minimum, you need a device to connect the PC to the network, and a slot on the PC into which you'll plug the device. With those items, you can create a network of wireless computers. If you'd like to connect the wireless network to a wired network, you use either a gateway machine (containing one wired card and one wireless card) or a local bridge that includes both a transceiver and a connection to the wired network. You can see some of the wireless LAN possibilities in figure 29.4.

Figure 29.4 Wireless LAN configurations can appear in many different forms.

To connect dial-in wireless clients to the network, you'll need a wireless modem (many fit into a PC-Card slot) and some kind of cellular or packet-switching wireless network service. You can't start running a private wide-area wireless network any more than you can get a bulldozer and start installing cable across the county; to provide WAN service you need right-of-way, whether it's to run cable or to utilize a given frequency. Public wireless networks are available in most urban areas in the United States.

Wide-Area RF Protocols

Over distances longer than a few hundred yards, some kind of protocol for setting up the connection and transmitting the data is necessary. Wireless communications can pass over either analog or digital signals, so we'll discuss the more common protocols that pertain to each.

Cellular Protocols

In the beginning, there was cellular.

Unless you've been living under a rock for the past decade or so, you're familiar with the concept of cellular telephones: small, wireless telephones with which the user can place calls within a certain area. Many urban areas are divided into hexagonal cells that contain a centrally-located transceiver to route calls to their destinations. You can see the arrangement in figure 29.5.

Figure 29.5 A cellular network can connect you to other cellular users, permit you to dial into your office, or pick up Internet e-mail on the fly.

Cellular communications work all right for voice (although there is a certain level of distortion), but isn't really suited for data transmittal, because it's slow (often limited to less than 1200 bps actual throughput) and because the noise and activity on a cellular connection are not good for data integrity or connection stability. It's got one big advantage over digital, however: the infrastructure is already in place. It may not be great, but it's there, and it works.

The lack of data support might not seem extremely important since cellular systems still carry mostly voice traffic, but data traffic is increasing in proportion to voice traffic in cellular systems. Voice usage of cellular systems isn't going away, but it's going to be coexisting more often with data usage. Thus, improving cellular's data handling is important.

MNP10

To alleviate the problems with cellular data transmissions, some developers began working on a protocol that could handle conditions unique to cellular transmissions, instead of trying to treat the connection like a faulty wire. In 1992, Microcom came up with a protocol named MNP10 that was based on the wired modem protocol MNP4. MNP10 had some error-checking capabilities and allowed for significantly increased throughput, but also had some problems. First, it was proprietary: to gain any benefit from the new protocol, the hardware at both ends of the connection had to support MNP10. Second, MNP10 perceived cellular events as noise, so the modems had to be resynchronized after every bleep on the connection. Cellular connections have a lot of bleeps, so this meant that MNP10 did not provide as significant an increase in speed as you'd expect. Modems spent a lot of time recovering from events, rather than transmitting data.

ETC

In 1993, AT&T/Paradyne developed a protocol named Enhanced Throughput Cellular (ETC). This protocol, based on the wired modem protocols V.42 and V.32bis, allows for speeds of up to 14.4 Kbps (before compression) and requires a modem supporting ETC on only one side of the connection (although AT&T recommended that ETC be present on both ends for best performance). Like MNP10, ETC sees cellular events as noise, requiring resynching time, but it has the advantage of being required only on one side of the connection.

CDPD

In 1994, IBM and others released a new way of sending data over an analog cellular system. This method, Cellular Digital Packet Data (CDPD) is a packet-switching overlay on top of the existing cellular system. In the overlay, data to be transmitted is placed in packets and then sent in the open bandwidth. It works much like wired packet-switching access protocols like frame relay and X.25, which are described in chapter 4, "LANs versus WANs." Like frame relay, CDPD is best suited for bursty transmissions rather than long, time-sensitive files. (Bursty transmissions are those in which the data comes in bursts, rather than in streams. If you're sending a graphic here and a text file there, those are bursty transmissions because they come and go quickly, and it doesn't matter what order the data arrive in so long as the ending file is in one piece. Time-sensitive transmissions, like voice and real-time video, are less suited to packet switching because they don't look or sound right unless the data in them arrives in a certain order and at a set pace.)

CDPD's similarity to frame relay is more than coincidental. In at least one CDPD network, frame relay is the access protocol used to transmit the messages. Recall from chapter 4, "LANs verus WANs," that frame relay is not a network typeóit's an access protocol to a network, either wired or wireless.

CDPD has a number of advantages:

CDPD won't replace older cellular designs completely, at least not yet. First, it's really meant for short, bursty traffic like e-mail and short text files. Longer files are more expensive to transmit via CDPD than traditional cellular transmissions, because it's billed on a per packet or per byte basis. Second, it's not widely available yet (as of late 1995) and it's not nationwide. To send data from one cellular region to another using CDPD, you must set up an address in the other region or send the data to a gateway mailbox to which the message's recipient has access. Third, if you leave your subscription area and enter another one, you may not be able to send dataóand the areas are pretty small because CDPD is designed to be very efficient with its bandwidth.

Even with its current limitations, CDPD might be the protocol that keeps cellular data communications from becoming obsolete. It certainly has the speed edge on the others, and the other bugs might get worked out with time. However, it has a long way to go before gaining real market credibility.

Enhanced Specialized Mobile Radio

Motorola and others offer a competitor to cellular services called enhanced specialized mobile radio (ESMR). This wide-area data service, built from underutilized fleet-management frequencies, is offered throughout many urban areas in the United States, and has the potential to become a unified national service. It was formally approved as a wide-area transmission protocol by the FCC in 1991.

ESMR takes low-bandwidth frequencies, digitizes them and uses TDMA to create as many as six channels in each frequency. Digitizing data makes it possible to reduce its bandwidth use, and time-slot channels should allow more than one subscriber to simultaneously use the same frequency without noticing any degradation in service. The equipment required is a ESMR-compatible transmitter and receiver, available from the vendor.

Although its frequencies have lower bandwidth than those used by cellular communications, ESMR can compete because of its significantly lower startup costs, and effective use of the bandwidth it has. If ESMR makes it as a national service, and is marketed well, it could offer stiff competition to the entrenched cellular services.

PCS

Although it's not out at the time of this writing (late 1995), personal communications services (PCS) promises a great deal to wireless networkers. It's going to take some time (the service isn't scheduled to be offered until 1996 or 1997) to see if it can live up to those promises.

Similar in structure to cellular services because of its physical organization, PCS has the following characteristics that distinguish it from analog cellular systems:

In February 1995, the FCC allocated 50MHz of noncontiguous spectrum to private sector use. Three of the bands within this 50MHz were destined for PCS. The idea is that the PCS frequencies will be isolated in very small cells, and the frequencies used for transmission will be used as efficiently as possible.

High-Speed WAN Connections

Although signals sent in the HF and VHF bands aren't high-intensity enough to make good inter-building transmitters, those in the SHF band (3ñ30GHz), or microwaves, are sufficiently high-intensity to provide a reasonable level of bandwidth. On a more-or-less line-of-sight basis (that is, the transmitter and receiver must be visible to each other), you can transmit data at high speeds between buildings or facilities, without having to install a cable.

Because microwave must be line-of sight, the transmitters and receivers are often not on the buildings that need to be connected but are instead on towers on mountainsides or high hills and then connected via a high-speed cable to the main network. The higher the tower is off the ground, the wider its range.

Microwave is limited in a few ways. First, the microwaves are bad for humans (the reason why you don't see many warning signs about microwave ovens and pacemakers anymore is because the ovens are better shielded than they used to be, not because the waves are less dangerous). Second, you must apply to the FCC for a license to transmit along a certain SHF frequency, and the license requires both time and money. Third, the equipment is expensive. However, if you really need a long-range secure wireless connection, then microwave is a good bet.

Limitations of RF Wireless Networks

RF networks have some limitations, of course. First, they frankly don't offer the performance of a wired connection. Unless there's a physical reason why you can't run cable in your office building, desktop machines benefit from a wired connection. Wired networks generally are faster, cheaper (not surprisingly, a transceiver is much more expensive than a network card), and since desktop machines aren't often moved, the lack of mobility inherent in wired communications isn't any drawback. Most of the time, if you must move a desktop PC, you disconnect it from the network, lug it to its new location, and plug it back into the network without any major hassle. The minute it takes to reconnect a computer to the network hardly justifies the slower speeds and greater expense of a wireless connection. In general, if you don't have portable PCs (notebooks and laptops) that require network access, wireless technology really is not for you.

Second, even office networks composed mainly of portable PCs can have wireless problems. RF signals can be stopped by obstacles such as thick concrete walls or iron girders. Even without physical barriers, distance plays a part in wireless office networking. About 300 to 500 feet is the distance limit within most offices, extending to around 800 to 1,000 feet outdoors or in very open office environments. Wireless networks work best in open areas without a lot of walls to block the signal. The more walls you have in the way, the shorter distance the signal will travel.

Third, too much wireless traffic can cause congestion on the network. RF technology uses a limited band for transmission, and if more clients attempt to use that band than there's capacity for, the signals can interfere with each other. Signals from other sources on the same frequencies can also interfere with wireless network transmissions. Bandwidth-sharing technology like TDMA and CDMA can somewhat alleviate the traffic situation, but it's still a point worth considering. For this reason, many transmissions used in urban areas and within offices use lower wattage than the FCC says they can, to reduce the amount of crosstalk between transmissions.

Infrared Wireless Networks

Infrared (IR) technology is not new (it's part of your television's remote control device, for example) but it's relatively new to wireless computing. This section discusses what infrared signals are and how they're coming to be used in wireless networking.

What Are Infrared Signals?

Infrared frequencies are those found at the extreme upper end of the visible electromagnetic spectrum. They have extremely high bandwidth and a correspondingly short range, as discussed earlier in this chapter. For the most part, infrared signals are limited to applications where the receiver and transmitter are directly in sight of each other, as IR signals can't pass through walls or most other fairly solid obstructions. Brief interruptions, like people walking in front of the signal, won't break an IR connection, but protracted interruptions will.

The short range of infrared signals means that IR, as a technology, is currently more generic than RF. For the most part, an infrared signal is an infrared signal; they're kept from interfering with each other by their low diffusion and short range.

IR Applications

The use of IR signals in networking is still in its infancy, but here are a few of the current applications:

The following sections explore these applications in detail.

Laptop Networking

Although infrared transmissions don't mix well with the rabbit-warren layout of many offices, mobile computing is used in more places than the office. For example, a college computer class can use laptop computers networked with infrared transceivers, allowing students in the open area of the classroom to network, but keeping the signals from leaving the room. Thus, the members of the class can interact with each other and the professor. In addition, with the right software in place, students can get help from the professor if they run into problemsóthe professor can bring up their screens on her computer and see what the problem is.

Wireless Printing

Hewlett-Packard recently released the LaserJet 5P, the first printer with a built-in infrared port. A PC with an infrared transceiver can send print jobs to the printer via this port without being cabled to the printer. The printer and PC must be located fairly close together (about three feet according to HP, but about six feet in real-world testing) and the signal path unobstructed. The PC and printer must be on stable platforms to maintain the connection. Other than that, the connection works like a wired connection, even matching the speed of a printer cable.

If you don't have an HP 5P, you still can make your printer wireless with a device named JetEye, offered by Extended Systems. Via a six-foot cord, the receiving device plugs into the parallel port (meaning that only IR devices can access the printer when you've got it set up for wireless networking). The device works up to four feet from a notebook under optimal conditions.

Connecting Buildings

Although IR is not currently supported for most wide-area networking, one device, SilCom's FreeSpace, permits buildings close to each other (up to about 1,000 feet apart) to be connected at native LAN speeds. Ethernet, full duplex Ethernet (20 Mbps), and both token-ring LANs are supported currently, with ATM and Fast Ethernet support planned for the near future.

Why not use a fiber to connect the sites? Well, it's not always feasible or cost-effective to do so. First, if there's a road between the buildings to be connected, you can't run a cable across the road or dig up the road surface to run it beneath. Second, the telephone company might not run the fiber for you: in the United States, at any rate, most WAN vendors have stopped offering dark fiber since the FCC told them that they no longer have to do so. Third, if you need something that can be set up fairly quickly, a solution that doesn't involve running cables is faster to install than one that does.

Why not microwave? Cost and convenience. The IR device has a shorter range, but it's cheaper than microwave and doesn't require an FCC permit to use a certain signal.

A wireless building connection works this way: three devices, one on each building and another between the buildings, handle the IR transmission. The device between the buildings sends signals to, and picks up signals from, the transceivers on the sides of the buildings. Within each building, a fiber cable connects the transceivers to a bridging device that connects the fiber to the network. As long as the bridging device can support both fiber on the transceiver side, and Ethernet or Token Ring (as appropriate) on the LAN side, it doesn't matter what kind you use. The whole setup resembles the one shown in figure 29.6.

Figure 29.6 The components of an inter-building IR connection are not complex.

If you want to connect two buildings within sight of each other that are not very far away, and a fiber connection is either too expensive or impossible, an IR solution might suit your needs.

Consider the following advantages of using an IR connection between buildings:

Although infrared has limited applications, it looks capable of performing well within its sphere.

Disadvantages of IR

The biggest problem with IR wireless networking is a lack of support from the PC end. As of late 1995, not very many PCsóportable or notóhave infrared ports (IBM's ThinkPad is one example of a notebook that does). Additionally, if you have an IR port, you need software supporting it so that connected devices can use the wireless connection. Moreover, no infrared bridges seem to be on the market, so you must take the short range of IR connections into consideration when determining whether it works for your needs.

There are other considerations as well, including the environmental restrictions of walls and the possibility of interference from other light sourcesóbut neither of these concerns is insurmountable. If an IR device can't pass signals through walls, then you can use fiber to connect it to a device that can. In fact, for some applications, the physical restrictions that walls impose are more an asset than a liability, since that isolation increases security and potential frequency reuse. Once there is more support for the technology, you can expect infrared transmissions to be increasingly prevalent in wireless networking applications, within the boundaries of their usefulness.

A Comparison of RF and IR Signaling

Because RF and IR signals are dissimilar enough to compete significantly with each other, it's worthwhile to recap their relative advantages so that you can decide which one might best suit your needs.

RF Signaling

RF's main advantage is that the technology is well entrenched. Because the first forms of wireless networking used RF signals, a good deal of development work has gone into extending the capabilities of RF transmissions. It's not difficult to find RF-compatible networking solutions, in the form of either wireless modems or transceivers for PCs.

Second, RF signals are comparatively durable. Despite the fact that higher frequencies are more affected by interference and blocking than lower ones, any RF signal deals with interference better than a microwave or infrared signal. Infrared transmissions require line-of-sight between the transmitter and the receiver. RF signals do not, although they do perform better with fewer obstacles.

Safety is also an issue. RF transmissions can be used both indoors and outdoors, since at this point no harmful effects are associated with them. If high-powered enough, IR signals can be dangerous indoors (or anywhere that they're easily accessible). The transmitting beam of a powerful infrared deviceóor even a laser pointerócan damage your retina if you look directly into it.

Of course, not all infrared light is dangerous; for instance, biometric security devices that scan your retina use infrared light. The rays are not dangerous at that intensity (the devices are not especially reliable, either).

IR Signaling

For all of RF's entrenched position, IR could offer it some competition if the industry supports IR. First, it's high-speedóable to keep up with wired LANs and more. This isn't very important in places where wired LANs are possible, but when you want to connect LANs in a situation where you can't just tack another segment on the network (such as connecting a LAN to another LAN in a neighboring building), IR can make the connection with no loss of speed.

Second, IR doesn't have the bleed-over problem of RF signals, which can interfere with each other if the frequencies used are too close together. IR signals are tight beams with a short reach, so they tend to keep to themselves.

Third, because of the concentrated beam they use, IR signals are more secure than some other types of wireless transmissions (especially non-encoded ones that use the same frequency for the life of the connection). They're not totally secureónothing isóbut tapping them would require getting very close to a short signal, which is hard to do unobtrusively.

The Future of Wireless

Don't expect wireless networking to replace wired networking; it will work in tandem with it. If you've got a wired LAN, then there's no impetus for you to run out and replace the cabling and network cards you've already got with transceivers. You'd be spending a lot of money to reduce the reliability and performance of your network.

As you've seen in this chapter, however, there are real applications for combining wired and wireless networks. You can connect two wired LANs in adjoining buildings using an infrared transmitter, or permit a mobile employee to dial into your office network without having to search for a telephone jack. These capabilities, and many others, have the potential to extend computer networking to areas where it would have been impossible in previous years.

Summary

Wireless technology is not a monolith; it varies in form and application from sending files to a printer to connecting buildings miles apart. Wireless takes two main forms: radio-frequency signals, which use radio bands, and infrared, which uses light to transmit data. Although when wireless first appeared on the scene, some envisioned that it would replace wired technologies, it now appears more likely that wired and wireless networks will become more common, allowing networks to benefit from both technologies.