Cabled networks are far from dead, but wireless solutions can solve tricky installation problems.
Christopher Bailey
Black Box Corp.
Will wireless networks make cabled networks obsolete? Is it foolish to spend money on installing cable now when you could go directly to wireless?
It could happen, but not for a long time. Right now, wireless networks are simply not as good as wired ones. But wireless networking has become practical enough to make the performance tradeoff worthwhile when difficult installations come up. Every network installer should know a little about the kinds of applications that make wireless networking worthwhile.
Perhaps the best way to introduce what wireless networking can do is to point out first what it cannot do. Most kinds of wireless networking do not match the speed of wired networks. The cost of the equipment is much higher than that of wired local-area-network (LAN) equipment. Obstructions that cables would simply bend around leave many wireless networks helplessly blocked. And phenomena over which you have no control--like the weather--can dampen the performance of wireless links.
Why wireless?
So why would you go with a wireless solution if wired networks are faster, cheaper, and more reliable? Because there are places where it is simply impossible or impractical to run cable.
Suppose you need to link two buildings separated by a busy boulevard. You can`t dig up the boulevard to run cables under it. High-speed leased services from the telephone company are prohibitively expensive and probably much too slow. The obvious solution is to link the two buildings without cables. In this case, you can put up with the minor disadvantages of a wireless link because the wired alternatives are worse.
Another case in which wireless is a good alternative is when installing a network in a protected historic building. Local ordinances are usually very strict: Every historic detail of the interior is untouchable. A wireless network may be the only way to create a network at all.
Trade shows offer another opportunity for wireless. Suppose you have a trade-show display to set up. You need a network you can plug in quickly and tear down just as quickly. A tangle of cables might take hours to sort out and would get in the way, but a wireless network sets up in a few minutes--and because of the small number of stations, the lower bandwidth is less of a problem.
Even in wired networks, wireless segments can be useful. A notebook computer wired into the network is no longer portable, but a wireless Ethernet card sets it free again.
The main use for wireless networking, then, is solving problems that make cabling difficult or impossible. As the technology stands now, cabled networks are faster and cheaper in ordinary installations. The extraordinary installations are the ones that make wireless networking attractive.
Different problems, different solutions
There is no one-for-all answer to the question of which wireless network is best. Just as with wired networks, different installations pose different problems, and different equipment will solve them. The most important question to ask is: What exactly is the problem you need to solve?
If the problem is that a few users need mobile access, for example, it would be foolish to pay for mobile access for everyone. It would cost far too much money, and it would impose a performance penalty that most of the network users simply should not be paying. A reasonable approach is to use cable for most of the network, with mobile users served by a few wireless access points connected to the main network.
Likewise, if the problem is a hard-to-cross gap between buildings, a single infrared laser link between two wired networks might be the solution. On the other hand, if the problem is a temporary installation in a trade show, where it would be ugly and dangerous to have cables crossing the floor, a complete wireless network might be the best option.
Sometimes the problem is simply financial. Larry Bertoncello, a wireless LAN expert with the Black Box Corp., says high monthly charges for wired internetworking often drive people to wireless. "People want to get away from the monthly charge for the phone lines, or they can`t get a T1 [1.554-megabit-per-second] line or 56-kilobit-per-second line," he says, "so they can buy wireless equipment and just have the one-time fee. It pays for itself after so many months."
Bertoncello`s job is helping people decide whether and how to use wireless networking. He hears about all kinds of installations, but the one thing they have in common is that it would be difficult to accomplish the same objectives with cables. The applications he deals with show just how diverse the wireless market already is:
Campus environments--People often use wireless because they have a campus-style environment with a remote office on the same property as the main building. They don`t want to have to run fiber or twisted-pair cable to it, so they use wireless.
Shipping--Data can be transferred to and from ships on the waterways.
Schools--Many schools are starting to put wireless in temporary classroom trailers so they can get their LAN traffic out to the trailers.
Manufacturing--Chrysler Corp. has installed wireless networking equipment around its test track and in the cars to transmit testing information from the cars back to the computer.
Medical--Hospitals often put wireless equipment on a cart that is moved from room to room when doing patient tests. The equipment sends data back to the main system at the nurses` station.
Security--In a certain oil field where oil was getting stolen, the owner installed a sensor. When the oil levels started dropping, the sensor sent a wireless signal back to the office, and security was dispatched.
In Bertoncello`s experience, many different problems can be solved by some kind of wireless networking. But whatever the problem is, identifying it gives you the first clue to solving it. Once you know what you plan to use wireless transmission for, you already have a fairly good idea of what technology you need.
Radio vs. optical transmission
Two basic kinds of transmission dominate the wireless market: radio and optical. Both technologies are useful but in different applications.
Optical transmission usually relies on lasers. For most optical links to work, you have to have a clear line of sight between the two ends of the link. (Diffuse infrared technology, still relatively unusual, does not require a direct line of sight, but it does require that the two ends of the link be in the same room.) Any obstruction--even a heavy morning fog--will simply cut off optical transmission. Because of that limitation, optical transmission works best for short-distance, point-to-point links.
The advantage of some optical systems is that they can reach much higher speeds than radio systems--even up to Fast Ethernet or Asynchronous Transfer Mode speeds in some cases. If there is a clear path between the two sites to be connected, a laser system can be the best choice to bridge the gap between two wired networks.
Radio transmissions are not limited to line-of-sight; within certain broad parameters, radio waves can go through walls and around corners. Unlike lasers, radio waves can propagate in all directions at once. They are subject to interference, but using spread-spectrum technology eliminates a lot of the interference.
Completely wireless networks and wireless access systems for mobile users usually rely on radio technology. In a network of any size, the access points are distributed so they form overlapping "cells" just like the cells in cellular telephone service. Sometimes, when multiple cells overlap, it is possible to combine the access points for more bandwidth. In most of these cellular schemes, "roaming" is possible--that is, a mobile user can move from one cell to the next and stay connected to the network.
The main problem with radio transmission is that it usually is not very fast. Equipment that follows the IEEE 802.11 standard, for example, is ordinarily limited to 3 Mbits/sec--slow compared to Ethernet and glacial compared to Fast Ethernet. Proprietary systems commonly reach the standard Ethernet speed of 10 Mbits/sec, which is still not very fast by today`s standards.
Standard or proprietary?
The advantage of going with accepted standards is obvious. You know the equipment you buy now will work with other equipment that follows the standards, even if the company you buy it from goes out of business.
But there are some disadvantages to insisting on the standards, too. The standards bodies move at their own deliberate pace. By the time a standard is official, technology may have marched right past it. Consider the IEEE 802.11 standard, for example. It`s meant to be the standard for cellular wireless networks in average offices, but under ordinary circumstances, the best throughput you can expect from equipment that sticks to the standard is 3 Mbits/sec. That throughput may be good enough for light-duty networks in small offices, but it is certainly slow by the standards of most wired networks. On the other hand, proprietary systems with 10-Mbit/sec throughput are common these days. The technology has raced past the standard.
Of course, newer and faster standards are always in the works. Will today`s proprietary equipment meet tomorrow`s standard? The makers of proprietary equipment often try to guess what the eventual standard will be, but experience shows that guesses like that are more often wrong than right. With luck, today`s equipment might be software-upgradable to the new standard but it would be foolish to assume that without a written guarantee from the manufacturer.
In the world of wireless, then, the choice often comes down to a maddening dilemma. Choosing a proprietary system means sacrificing interoperability. Choosing a standards-based system means sacrificing throughput.
In fact, the choice can be even more difficult. Even sticking strictly to the 802.11 standard doesn`t guarantee compatibility. "What people don`t understand is that there are two sides to that spec," explains Bertoncello. "You`ve got frequency-hopping and direct-sequence. So people ask, `Do these units support 802.11?` Then they get them and they still don`t work. Not only do you have to ask whether it supports 802.11, but you also should ask, `Does the equipment do frequency-hopping or direct-sequence?` "
In the end, the application will determine whether a proprietary system or a strictly standards-based system is the best choice. Installations where the most important consideration is throughput will probably need high-speed proprietary systems--but you will run the small risk of being stuck with a network you can`t expand or upgrade. On the other hand, in growing networks where new stations are constantly being added, sticking with the standards makes the growth much more manageable--but you will have to be satisfied with slower performance.
In short, dogmatically insisting on the standard in every case is probably not the best solution to the problem, but neither is dogmatically insisting on the fastest proprietary system. Answer the question for each installation individually.
Too good to ignore
Technology is always racing forward, and it`s difficult to say what will be coming even a year from now. But networking without wires, in spite of the current limitations of the technology, is simply too good an idea to be ignored. We can predict with confidence that the technology will continue to improve, and wireless networking will become more practical in more applications. If you`re a network installer, your customers will have heard of wireless networking. Some of them are bound to start asking what the possibilities are. Listen to the questions they ask. If they`re asking about wireless networking, it`s probably because they have a problem that they think might best be solved that way. You might be able to suggest a cabled alternative that would work as well or better. But some problems do require a wireless solution. Even if you can`t provide that solution yourself, you need to know enough about the subject to make educated recommendations.
Laser transmission can be a good choice for short, point-to-point links. With a unit like that at each end of the connection, transmission speeds can reach 155 Mbits/sec.
Radio-wave-based wireless networks make use of access points, which create "cells" of network reception. Where cells overlap, mobile users can "roam" from one cell to another, and higher bandwidth may be possible.
Avoiding T1 costs is a popular motive for installing wireless systems. In the figure, a one-time investment in wireless equipment links several buildings in different parts of a city.
For More Information...
For information about wireless technologies and products, visit the Website of the Wireless Local Area Network Alliance at www.wlana.com. The site also includes white papers and a survey on actual wireless implementations.
Christopher Bailey is a technical editor at Black Box Corp. (Lawrence, PA--www.blackbox.com), a provider of computer-communications and networking products and services.