 |
||||
 |
||||
 |
||||
 |
||||
 |
||||
 |
||||
|
||||
 |
A Primer on Wireless Networking Essentials
by J. Eric Smith, MCSE, CNE, CCNP, CISSP
So you've finally done it: you've committed your organization to deploying an 802.11-based wireless network. If you're like most people, this decision was not arrived at lightly, with security and usability concerns weighing heavily against the needs of the more mobile, flexible workforce of today. But now it's in the budget, and you're on the hook to deliver. Only one question remains: exactly which 802.11 standard will you deploy?
The fact that there are multiple types of 802.11 networks can come as a rude surprise to the wirelessly uninitiated. Unlike the tried and true Ethernet with its well-known simplicity and ubiquity, wireless networks have followed a more tortuous, divergent path to their current incarnations. The result is a dizzying array of letters following the 802.11 moniker, such as 802.11a or 802.11g. Each one denotes a specific subset of the 802.11 standard, and understanding what each letter brings to your network makes the difference between deploying a marginally-capable wireless infrastructure and one that will serve your organization for years to come.
Key factors to understand when designing a wireless networking solution are as follows:
Frequency band - This is the portion of the electromagnetic spectrum where the device operates. For networks in the United States, this spectrum is governed by the Federal Communications Commission (FCC). The various 802.11 standards all operate in what is called the "unlicensed" frequency spectrum, a spectrum that covers portions of the 900-megahertz band, the 2.4-gigahertz band, and the 5-gigahertz band. These bands are called unlicensed because anyone is allowed to transmit within them without the need for an FCC license. While this makes for a fantastically diverse array of products, it also means that anyone can potentially interfere with your 802.11 network. Also, certain frequencies penetrate structure better than others. Other frequencies are better at high data rates, while still others are easily interfered with by cordless phones, microwave ovens, and baby monitors. Proper spectrum selection is crucial in order to get the best range and least interference for your intended application.
Frequency bands are divided into discrete slices known as "channels," with each channel being assigned a number. An important wireless concept to understand is how certain channels overlap other channels, restricting usage to far fewer channels than you might think. Access points operating on the same channel will interfere with one another and prevent proper operation of either one.
Data rate - Often expressed in terms of megabits (mbit) per second, quoted data rates should be taken with a large grain -- or perhaps a small boulder -- of salt. Data rates shown by most vendors are the theoretical maximums achieved in laboratory conditions -- something you'll never find in a real-world environment. Most effective data rates are less than half the quoted rates -- an important thing to consider if your intended application is bandwidth-heavy. Data rates also decay rapidly as devices move further from the access point, making proper coverage planning essential.
Density- Unlike contemporary Ethernet switches, wireless bandwidth is shared among all devices associated with the same access point. This puts a reasonable maximum on the number of systems that can be used in a given area covered by a single access point. If your intended usage is a high-density environment such as a call center, selection of the proper 802.11 technology is crucial for success.
Range - Range depends heavily on three things: transmit power, the aforementioned frequency band, and the structure within which you deploy the wireless network. Of these three items, only one -- the frequency band -- is within your control. The FCC controls the maximum amount of transmit power, but careful selection of the proper antenna and access point will allow you to make the most of it. As noted above, certain frequencies are better at going through structure than others. Proper band selection is thus crucial to getting the range you need in order to get the coverage you want. But with greater range comes a greater possibility of interference from other radio sources, making proper access point placement the deciding factor between success and failure.
Features - The unique security issues surrounding wireless networks have spawned a wide array of features in contemporary wireless products. Enterprise-class devices will often feature support for a wide variety of encryption protocols such as WPA and WPA2, authentication mechanisms such as 802.1x, and even exotic features such as the ability to triangulate any wireless device down to an area of one meter -- your very own in-house GPS system, if you like. With the base features of 802.11 products rapidly reaching commodity status, many vendors are offering distinctive feature sets to differentiate themselves from their competition.
With these criteria in mind, let's visit some of the common 802.11-based systems you're likely considering right now. We’ll also detail the pros and cons of each.
802.11: An Uncertain Beginning
This is the base wireless standard that debuted ca. 1999. It featured a 2-megabit (mbit) data rate and operated in the 900-megahertz (MHz) frequency band. Being the first iteration of a standard, it also had vendor interoperability problems, as different vendors "interpreted" the 802.11 specification slightly differently. The 900MHz frequency band also became quite cluttered with first-generation analog cordless phones, baby monitors, and garage-door openers which were notorious for their inability to play well with wireless networks.
This standard has essentially been abandoned today, its data rate being far too anemic for most uses. No vendors are actively pushing any products that are restricted to the original 802.11 specification.
At a Glance:
|
802.11b: Wireless Hits Its Stride
Following closely after the original 802.11 -- and skipping over the 802.11a standard which was still in the works -- 802.11b also debuted ca. 1999. The IEEE committee learned from the mistakes of the preceding standard and made this one much more rigid. The result was the first real wireless standard that offered inter-vendor compatibility. With buyers no longer being tied to a single vendor solution, the wireless market exploded. To date, there are more 802.11b devices deployed than any other, making it the de facto standard you're likely to find anywhere you go.
802.11b operates in the 2.4-gigahertz (GHz) frequency band and has a quoted data rate of 11mbit. Its structure penetration is only slightly inferior to the 900MHz frequency band used in the original 802.11 specification. Although the frequency band is divided into 14 channels, North American users are restricted to channels 1 through 11 (Japanese users can use the extra channels). However, due to the way the channels are allocated, only channels 1, 6, and 11 can actually be used in proximity to one another. As noted above, planning channel allocation goes hand in hand with planning coverage and density. Too many access points in too small of a space will result in overlapping channels, killing all access points in the area. This puts a practical upper limit to how many people can reasonably use any wireless network in a small space.
At a Glance:
|
802.11a: The Alternative That Didn't Catch On
Proving again that engineers can't count (or at least can't sequence their alphabets properly), the 802.11a standard came out after the 802.11b standard. In almost every respect, 802.11a is superior to 802.11b, and that has largely been its downfall: 802.11a engineers sought to correct the many shortcomings of 802.11b, and in the process they created a standard so advanced it was completely incompatible with 802.11b.
It didn't help 802.11a that 802.11b had been a smashing success, either. 802.11a uses the 5GHz frequency band and thus requires a completely different kind of antenna than the 2.4GHz 802.11b devices. In the computing world, economies of scale trump all, and 802.11a devices were debuting at prices three or four times higher than the existing 802.11b devices. And since 802.11a's 5GHz frequency band is less effective at penetrating structure, range decreases by as much as 50% compared with 802.11b devices. To date, 802.11a has sold very poorly. It is virtually non-existent in the consumer market and is sometimes difficult to find even in the professional market.
But this isn't to say 802.11a isn't a fantastic wireless variant. By breaking with 802.11b, engineers were able to wring a 54mbit data rate out of the new standard, nearly five times faster than 802.11b. Furthermore, 802.11a was endowed with a much larger number of available channels, reducing the chances of interference from other 802.11a devices. And moving to the 5GHz frequency band also meant moving away from all the potential interference sources currently plaguing the 2.4GHz frequency band. These are not insubstantial advantages, especially in environments where low interference -- and thus higher reliability -- is paramount.
At a Glance:
|
802.11g: The Reigning Champion
Shortly after the apparent failure of 802.11a, engineers decided to do something a little less radical. Picking and choosing the best aspects of 802.11a and 802.11b, they came up with 802.11g. This variant is fully backwards compatible with all 802.11b devices -- an advantage that gives it an immediate installed usage base of millions of devices. 802.11g devices are currently the best selling products.
In order to remain compatible with 802.11b, 802.11g sticks with the 2.4GHz frequency band. But by using the more advanced encoding mechanism of 802.11a, the data rate goes up to 54mbit. But with backwards compatibility comes some unwanted baggage.
802.11g shares 802.11b's limitation to only 11 channels, out of which only three are usable. Worse, any 802.11b device operating near any 802.11g device will degrade the 802.11g device's data rate somewhat. And since 802.11g also operates in the 2.4GHz frequency band, it is susceptible to all the same interference sources as 802.11b.
At a Glance:
|
Design and Deployment Challenges
As you might have guessed by now, any wireless network is designed around a series of compromises. There is no such thing as a single, perfect, one-size-fits-all design, which means each of the following criteria must be answered and weighed against each other. They are: coverage,density, and intended usage.
Coverage Of the three, coverage is the easiest in practice to achieve by itself. One must merely saturate one’s building with signal in order to achieve signal coverage. Unfortunately, such a deployment is almost certainly not going to work well because it neglects the second criteria, namely density. Also, coverage can be considered a two-edged sword, as adding more access points leads to greater signal leakage outside the areas you control as well as offering more opportunities for something to interfere with your network.
Density is derived from the number of devices you wish to be on a wireless network in a given area divided by the maximum number of devices a single access point will support. As a general rule of thumb, most experts will say that you want no more than 20 users per access point, although this figure is heavily dependent upon how those users are using the network. If you consider that the average 802.11b or 802.11g network will only allow three access points to be in close proximity to one another, you'll quickly find that it's difficult to support more than 60 wireless users in a confined space. For office environments this is rarely a problem, but lecture halls, theaters, and even large conference rooms can be exceptionally difficult to accommodate without a proper design.
Intended usage is the last criterion but perhaps one of the most important ones. Infrequent, low-priority wireless usage -- for Internet browsing by guests, for instance -- is much less demanding than a bandwidth- and latency-sensitive service like Voice-over-IP (VoIP). Whereas the former usage might support fifty users on one access point quite easily, the latter might have difficulty supporting three or four simultaneous calls with one data user. This criterion is also the one most likely to change over the life of the wireless infrastructure. This means that if any criterion requires in-depth planning for the future, this is it.
In summation, the proper wireless network will have coverage where you want it, but only where you want it and not spilling out into the parking lot for hackers to prey upon. It will accommodate the density figures you require, but not be too dense or your network will actually interfere with itself. Lastly, it will provide you the ability to do what you want it to do with room for future growth, but will not be so over-the-top that it will bust your budget. To say that this is a fine balancing act is an understatement.
What Does the Future Hold for Wireless Alphabet Soup?
While 802.11a, b, and g may be the letters most frequently used in wireless discussions, there is quite literally an entire alphabet of lesser-known variants and extensions to 802.11. While some of these are esoteric or submerged within 802.11 itself, others are going to be become as well known as -- or perhaps better known than -- their contemporary brethren over the next year. Knowledge of them and how they will affect the wireless landscape is essential to planning a wireless network today. They are:
802.11i - This currently-ratified standard is perhaps better known by its acronym WPA2, or "Wi-Fi Protected Access II." This is the latest encryption standard to be offered on wireless networks and it works with 802.11a, b, or g. The prior wireless encryption standard was the widely-disparaged Wired Equivalency Protocol (WEP). WEP gave wireless in general a bad name in the security world because it has been laughably easy to circumvent for several years. 802.11i, however, is based upon an encryption standard certified by the U.S. government for use in agencies where secrecy is paramount. The encryption cipher is known as the Advanced Encryption Standard (AES) and is widely considered to be one of the strongest in the world. No known vulnerability exists in it. Any solution you are considering should support this standard. If it doesn't, you should be looking elsewhere.
802.11n - The letter "n" might as well stand for "next big thing" given its sweeping implications. 802.11n seeks to succeed where 802.11a failed, namely at producing a radically-advanced 802.11 variant that is still somewhat backwards compatible with existing wireless networks. The goals of 802.11n are to increase the indoor range from tens to hundreds of feet while simultaneously increasing the data rate. Exactly what the data rate will be is still a matter of conjecture because the standard has yet to be ratified, but the IEEE committee has publicly stated that it has a target rate of 100 megabits of usable -- not theoretical -- bandwidth. If this target is met, it would represent a five-fold increase over the real-world data rates achievable with either 802.11a or 802.11g devices. This increase, coupled with an amazing increase in range, is sure to make 802.11n the technology that will replace all existing 802.11 networks when it becomes available in late 2006 or early 2007.
802.11r - Also known as "fast roaming," 802.11r seeks to address a problem between wireless networks and VoIP usage. Wireless devices, being wireless, have a tendency to be quite mobile. On a wireless network, this means the device will eventually leave the coverage area of one access point and enter the coverage area of another, much as a driver on a cell phone will "roam" from one cell tower to another while moving down the highway. When this roaming happens, the wireless device must re-establish its existing session with the new access point, a process than can take upwards of 100 milliseconds. While this presents no problem for a data-only user, voice users will suffer from interruptions or even dropped calls. 802.11r, once ratified, will standardize a handoff mechanism between access points, allowing roaming voice users to experience uninterrupted calls.
802.11e – This is another extension to 802.11 aimed at fixing some of the shortcomings associated with wireless voice calls. Currently, all types of traffic on a wireless network are treated equally. While a data-only user surfing the Internet will never notice a delay of a few hundred milliseconds, a voice caller will experience a significant interruption similar to a bad cell phone call. One way around this would be to have the network assign a higher priority to voice traffic over data traffic, a technology known as Quality of Service (QoS). Thus, when data traffic threatens to impede voice traffic, the voice traffic will automatically be given a higher priority. Wired Ethernet networks have had this technology for several years now. 802.11e will implement wireless QoS in much the same way. The lack of QoS is considered to be the last remaining obstacle to large-scale wireless VoIP deployments. You should expect an explosion in availability -- and demand -- of wireless VoIP products shortly after 802.11e is ratified.
Conclusion . . . for Now
Of all the technology infrastructure market segments, none is currently moving faster than wireless. Whereas wireless networks were barely visible five years ago, today it is difficult to find areas that aren't covered. The challenges of achieving adequate coverage have given way to the challenges of securing the network and accommodating an ever-growing list of required mission-critical uses. Tomorrow's applications will only bring additional challenges of greater complexity requiring even greater reliability. Current experience as well as an in-depth knowledge of the entire wireless market is essential to a successful deployment today.
If you are considering deploying a wireless network for your organization -- or if you've already committed to doing so and are exploring the various solutions -- the value of an experienced wireless consultant cannot be underestimated. Wireless networking is not merely networking without wires; it combines every challenge of wired networking with a whole new crop of design and implementation difficulties. Your organization's competitive edge as well as its security is on the line. Choose wisely.
.
|
A PDF downloadable version of this article can be found at http://www.ediltd.com/assets/wir elessprimer.pdf
About the Author:
|
