A Field Guide to Wireless LANs for Administrators and Power Users

A number of technology trends have been creating an environment in which wireless LAN (WLAN) technology can thrive. WLAN technology did not simply appear out of nowhere…the groundwork to prepare people for mobile computing has been underway for some time. Just as the Internet was able to explode because 1) PCs were abundant, 2) most PCs were connected by LANs, and 3) on the PC, the only thing required to become an Internet node was to install and configure the correct software (in particular, a TCP/IP stack and applications that could use it, such as a web browser). Had PCs not been commonplace, or had LANs not already been deployed, the Internet would not have been able to reach nearly as many people so quickly.

In the case of wireless (or, more generally, mobile) computing, the first enabler was the fact that laptops are getting faster, cheaper, and thus more prevalent. Another enabler was that people had become used to mobility due to the increasing popularity of cellular phones. WLANs enable laptop users to access the network as effortlessly as they can use a cellular phone to call home (provided they are in a location that has a deployed WLAN, of course). Finally, there are a number of application scenarios at home and at work for which the requirements ideally match the capabilities of WLAN technology.

Wireless and mobile communications are a feature that people increasingly have come to expect in relation to the user experience of computing. Mobility is becoming the norm, for the convenience and flexibility it affords. Eventually, it is likely that WLAN chips will find their way into most types of objects. There are a number of interesting vertical applications of WLAN technology, besides the obvious vertical application known as wireless Internet access.

For one example, in certain restaurants, orders are now directly entered at tableside by waiters using PDAs with WLAN capabilities.^[1] This is an excellent example of integration of appropriate technology to enable smoother business practices, since the PDA-based data entry gets the order into the system immediately, and the wait staff do not need to spend their valuable time doing data entry of the order (when they could be serving other customers).

^[1] The author knows of at least one restaurant that has implemented such a system. One wonders if the next step might be to offer wireless Internet access over the same infrastructure, thereby creating a revenue stream from the installed equipment.

Or in a hospital, one might imagine that doctors could use WLAN-equipped PDAs or laptops to pull up patient information at bedside, and be able to interact with that data much more fluidly than if it were a sheaf of papers. It is generally the case that hypertext capabilities are difficult to implement in nonelectronic documents. With regard to medical applications of WLANs, it turns out that certain types of hospital equipment may use the same frequency bands as certain WLAN technologies, so care would have to be taken to avoid interference. Such issues are avoidable if the WLAN technology is chosen such that it cannot interfere with the critical functions of various installed medical equipment; this could be accomplished by either physically separating the equipment or by using a different frequency so the two types of equipment cannot interfere with each other.

Chapter 2, A Brief History of Networking Standards, will cover the range of Physical layer choices for WLANs, and will make it clear what options exist for using two or more different types of communications equipment, via different operating frequencies, different modulations, or other parameters.

Rise of Laptops

Laptops are becoming the preferred type of personal computer, particularly for business users. When laptops were first brought to market, users lucky enough to have a laptop typically had both a desktop machine and a laptop, since laptops typically had less computing power than a comparably priced desktop machine and were only useful as an extension of their desktop PC, not as their primary machine. In addition to their initial high cost, early laptops tended to be heavy and not have very long battery life. These impediments tended to restrict laptops to applications in which mobility was mandatory, and for which the performance achievable on those early platforms was adequate. As time passed, laptops became lighter, their battery life was extended, and they became more powerful. Most importantly, they became cheaper so much so that for many users they are beginning to supplant the desktop machine.

Given that modern laptops have sufficient performance for most business applications, and that they have a significant edge in terms of portability, compared to desktop machines, their popularity has been rising, and the percentage of PCs sold that are laptops has been increasing^[2] as a result. It has become very common for a corporate user to only have a laptop. For home users, laptops are not as common, but the trend is that they are becoming more common, especially if one considers the term "laptop" in its broadest sense, and include new mobile PC product categories like tablet PCs that are designed around mobility from the start. In fact, one might observe that a tablet PC is relatively useless without a wireless infrastructure, so the WLAN has become an enabler for a new class of computing platform.

^[2] In fact, according to a news story on July 2^nd, 2003, the market research firm NPD Group has stated that in May 2003, revenues from the sale of laptop PCs exceeded revenues from non-laptop PCs for the first time. In terms of units, laptops still only account for 40% of the market, but since they are more expensive, they accounted for 54% of the market in terms of revenue in the month of May 2003. The trend is clear, however, and it would be reasonable to expect that laptops will soon lead the PC market in both revenue and unit terms.

Networking Became Pervasive

Until WLAN technology arrived on the scene, the only networking option for laptop users was Ethernet. There is nothing wrong with Ethernet. 10/100 Mbps Ethernet was (and is) ubiquitous in laptops, as it is in desktops. In the 2004 time frame, 10/100/1000 should begin to become "standard equipment" in new PCs, even in laptops. In corporate settings, Ethernet is the dominant networking technology, and a cubicle typically has from two to four network jacks. Part of the convenience of Ethernet derived from its use of the RJ-45^[3] connector, which is easy to use and is not easily dislodged once it is connected.

^[3] RJ-45 stands for Registered Jack-45, an 8-pin connector used commonly to connect computers onto an Ethernet-based LAN. RJ-45 connectors look similar to the ubiquitous RJ-11 connectors that are used as modular telephone jacks, but RJ-11 jacks are somewhat narrower, as they involve only 4- or 6-pin connectors.

However, using Ethernet requires that a user be near an available Ethernet hub or switch port to obtain network connectivity. This might sound like an obvious observation, and this is not an unreasonable requirement for a desktop PC. However, mobile users with laptop computers are always in search of a network (or telephone) jack from which they can obtain connectivity. While the laptop was portable, there was a certain lack of convenience in unplugging the Ethernet (or telephone) cable every time the user needed to move to a new location.

Ethernet was an acceptable (barely) "roaming" technology inside an office building, in which it was possible that all conference rooms and some common areas were equipped with network jacks and/or tabletop switches to which users could attach. Theoretically, providing wiring to conference rooms allowed a primitive form of roaming, but a room with four network jacks would be unable to provide network connectivity for five or more people, unless one brought his or her own hub or switch (and then the number of available power outlets in the room might become an issue!).

However successful Ethernet has been in corporate deployments, due to the widespread availability of network jacks, the typical home or apartment environment typically does not have an abundance of Ethernet jacks (or even phone jacks) in every room (in California, landlords are only required to provide a single phone jack in an apartment, and only newer homes have been pre-wired for networking). Deploying networking technology at home has, until the advent of WLAN technology, involved adding a wired infrastructure to serve the places that the homeowner thinks will need network connectivity.

Emergence of Cellular Phones

At the same time that laptops were emerging as the leading type of PC sold, cellular phones were becoming increasingly affordable, through a combination of competition and expansion to a larger user base. In fact, cellular phones have been in existence since the 1980s, but it is only in the recent past that cellular phones have become so prevalent as to be considered commonplace. In terms of public perception, it is now more unusual for someone to not have a cellular phone. So-called "mobile professionals" armed with laptops and cell phones are a common sight almost everywhere you go these days. Who hasn't seen people driving with one hand and taking notes on a dashboard-mounted notepad with the other while using a hands-free cell phone headset while shaving?

Cell phone networks are pretty good at delivering "good enough" voice communications, but a limiting issue with cell phones is that there is very limited bandwidth available for data services in today's existing cell phone networks. While it is possible to connect a laptop to a cell phone and dial out using a modem, as if the cell phone were a regular telephone line, the connection will rarely be faster than 9.6 kbps. Existing cell phone networks, even those that are digital, are optimized for voice, not data.

To remedy this situation, the world's telephone companies have proposed to enhance cell phones by upgrading to a new third-generation (often abbreviated as "3G") infrastructure, so that Internet-enabled wireless communications will be the norm, rather than the exception. In 3G networks, data performance will be much better (on the order of 100 kbps, or perhaps even higher). However, to get to this 3G "nirvana," an enormous investment in new infrastructure must be made. Moreover, the user base must adopt the new technology by purchasing new telephones capable of supporting Internet features.

Despite the existence of some agreements on 3G wireless standards, there have been few widespread deployments to date, and one reason may be that it is not certain how to cost-justify the investment. Will the new capabilities generate sufficient return on investment? Such questions were not asked in the late 1990s, when large sums of money were invested in infrastructure with little consideration of whether the investment would lead to increases in future income, and whether the possible increases were sufficient to justify making the investment. Now that the pre-3G investments are paying off, it seems like the business decision has been made to extract as much "bang for the buck" from the initial investment as possible, while doing trial 3G deployments to try to find out what new services users are willing to pay for. (It is amusing to note that the difficulties of getting 3G networks deployed have caused some vendors to start talking about 4G networks. Luckily, the evolution of cell phone network infrastructures is beyond the scope of this book.)

Even with the potential of 3G (or 4G, if you will) wireless technologies, such networks are not "LAN-like" and do not match the way people are used to interacting with networked content, especially with respect to billing. One of the major differences between surfing the Web at home or at work versus using advanced cell phone based data services is that the billing model is significantly different. In addition, to the author's mind, surfing the Web on a two-inch diagonal screen is of limited usefulness (and two inches is a large screen by cell phone standards).

Internet Billing: Wireless versus Wired

In the cell phone world, the providers seem to want to be able to create "value-added" services that they can charge extra for. There is no good reason that they should not behave this way. This is also how traditional telephone companies handle billing. Beyond your basic touch-tone telephone service, you can order additional services such as caller-ID, distinctive ringing, call waiting, voice mail, and so on. Typically, there is a per-use fee, or an "unlimited use" fee that is billed monthly.

There are probably some truly compelling mobile-oriented applications that phone companies will be able to offer as a data-dependent value-added service (things like "I'm at the following intersection…please make me a reservation at the nearest Chinese restaurant and tell me how to get there"). Value-added services like that would have a user interface that could be simple enough to put on a small screen.

Of course, the same application could be offered on a laptop via a Web browser, but how many people carry their heavy laptop with them everywhere^[4] ? General Web surfing and email^[5] are probably not going to be popular applications for cell phones, which are otherwise getting smaller and smaller.

^[4] Even if a practical one-pound laptop existed, it would still be over five times heavier than my cell phone…. Most people would agree that when it comes to mobile application platforms, lighter is better.

^[5] In some cases, email is a useful mobile application, such as the BlackBerry^® "über-pagers" by Research in Motion (RIM) Limited. A Blackberry is like a miniature laptop, in that it has a usable screen and a small keyboard, and some models even have limited Web-browsing capabilities (next-generation models will be able to be used as cell phones). These devices use a variety of pre-3G wireless data networks to maintain access in most major metropolitan areas, worldwide.

However, on a typical Internet connection, whether at dial-up or broadband (e.g., DSL or cable-modem) speeds, you usually get "All the Internet you can eat" (limited only by the speed of your access line) for a simple monthly fee. Once connected, Internet users are free to access whatever networked applications interest them, from the very popular World-Wide Web (WWW) to file transfer applications, to voice over IP (VoIP), and so on, only limited by the speed of their access method.

To an ISP, charging by the byte or by the connection would represent a vast amount of overhead in their billing systems, so Internet access is almost always a flat-rate service. Where optional fee-for-service "products" exist, they typically involve things like firewall service, or an extra IP address for nominal monthly charge(s), or some amount of online file storage, which may or may not be included in your monthly fee, on which you can store a Web site or email or what have you. Internet providers (that provide access via dial-up or via dedicated "always-on" technologies like DSL or cable modems) typically do not do their billing on a per-packet, per-message, or per-connection basis.

However, such charging models are not unheard of in the "Internet provider" context, as they are more likely to be seen in the "wireless Internet" offerings that are based on advanced cellular phones that double as Web browsers. The fact that the wireless Internet services are significantly bandwidth-constrained means that bandwidth is a scare resource; scarcity equates to value, so they need to set a price for every byte that a user consumes that reflects at least the basic value of the bandwidth consumed, plus some additional profit margin. If wireless bandwidth were significantly more plentiful, the pricing models would probably be different. As long as the bandwidth is scarce, it will be priced accordingly (i.e., it will be expensive!).

Wireless: Applications versus Devices

Today, wireless applications tend to be tied to a particular access device, with the laptop being the most general purpose, but also the heaviest, and devices like cell phones and pagers being on the other end of the scale, for very specific applications such as voice communications and interactive text messaging.

This situation of application-specific devices resembles the early days of mainframe computing, in which each application (or type of application) often required a dedicated terminal. It was not uncommon for someone to have two or more terminals on his or her desk, each devoted to just one type of application. Eventually, the "terminals" became software-based and all migrated into a general-purpose platform (initially, as special multi-window terminals, then the terminals evolved into programs running in the PC), sharing a single connection to the network.

At some point, the interactive terminal model itself (in which the applications, including the user interface and all the necessary support software ran on the mainframe) was replaced by networked applications using client/server paradigms, in which the user interface runs locally on the PC, and connects to a back-end server that provides access to some form of database or other networked service.

However, it remains to be seen whether there will ever be a "converged" wireless device that will satisfy the needs of a variety of applications. At this point, it seems more likely that application-specific devices will be the norm for some time to come, although some degree of convergence may occur as they all migrate to common delivery technologies based on the Internet Protocol (IP), perhaps over a future infrastructure based on 3G wireless technology, but also over other types of access networks, such as those based on WLANs, or perhaps wireless MANs.

It is also possible that application-specific wireless devices in close proximity to each other could pool their talents using short-range wireless technology such as Bluetooth_®. In such a scenario, users could assess their needs and assemble a set of wireless products (and their associated services) to match, obviating the need for an "all things to all people" common wireless device. For example, a PDA with a database of contact information could be queried by a cell phone so that when the cell phone rings, it can look up the calling party's number and display the caller's name instead of a telephone number. This approach is more flexible and modular, and theoretically more scalable, since it avoids the need to input the same information (i.e., the name and telephone number) into both the PDA and the cell phone.

As of early 2003, many late-model cell phones have Bluetooth capabilities, and once laptops and PDAs also have Bluetooth, this type of integration will just be a matter of software, since the connectivity will be available to enable all the devices to expose their unique services to the others, and allow the user's collection of devices to operate as a collective, seeming to be greater than the sum of its parts.