Several challenges associated with WLAN implementation could spell the difference between the success and failure of the project: shifting through standards, making sure that wireless coverage is adequate, preparing both the wired and wireless LAN for voice, tackling management and administration tasks, and planning for the prickly radio frequency (RF) environment. It's a long list, but you must understand them all to reap the benefits of a WLAN implementation.
There are no short cuts: These are very complex and potentially confusing networks which, if not designed, deployed, and managed correctly, can be a security and ROI nightmare. The enterprise unwilling to do the work necessary to design, install, and operate a WLAN properly is advised to bypass this technology altogether.
Confusion in the air
Hot technologies are always accompanied by a great deal of confusion. WLANs are perhaps even more confusing because four standards exist. The first goal is to figure out which WLAN standard—802.11, 802.11a, 802.11b, or 802.11g—is right for your enterprise’s needs.
The protocols break down as follows:
- 802.11: This standard offers transmission speeds of 1 or 2 Mbps in the 2.4-GHz range. It is considered antiquated.
- 802.11a: This standard offers transmission speeds of 54 Mbps in the 5-GHz frequency range.
- 802.11b: This standard offers transmission speeds of 11 Mbps in the 2.4-GHz frequency range and is associated with WiFi.
- 802.11g: This standard is not expected to become official until this summer, but it promises a transmission speed of 54 Mbps in the 2.4-GHz frequency range.
Even if you choose the best equipment from the choices available today, you run the risk of missing out on innovations that are just around the corner. At a minimum, experts like Mike McAndews, a director of business operations for Cisco, advise CIOs and their staffs to make sure that what they buy has a clear migration path into the future. “The emerging landscape suggests that a number of different standards will be supported in each device.” You are advised, however, to be careful until these devices emerge: Get a guarantee from the vendor that standards promulgated after the gear is purchased will be supported. Also check to see if the company is stable—if it's using proprietary gear and looks like it may go out of business, think twice about using it.
You should strive to create a working network that accounts for a wide number of variables, such as the number of users, their usage pattern, the physical location, and the standard that is being used. “The number one issue is the limits of performance and having the bandwidth to share,” said Dean Au, president and CEO of vendor AirMagnet. “If you have too many access points, you crowd the space and lose performance. If you have too few, there is not enough coverage.”
The use of airwaves means that these networks are more finicky than their wired equivalents. Indeed, even normal human activity has ramifications to coverage patterns. “The RF environment is dynamic,” Au said. “It never stays the same. Since it’s a shared media, it’s as simple as somebody rolling a metal handcart in a corridor that can change the dynamics.”
Unfortunately, the RF world is foreign to most enterprise CIOs. For that reason, you should carefully choose consultants or hire experts for WLAN implementation. Brice Clark, worldwide director of strategy and business planning for Hewlett-Packard’s HP ProCurve Networking Business, said, “There is a tradeoff between the size of the access point footprint and the number of users it can support. Careful analysis of usage patterns on an ongoing basis is important.”
Scalability is another challenge. “It’s one thing to put up an access point in your home,” McAndrews said. “It’s another challenge to create a wireless LAN for thousands in a network. You have to manage upgrades, and you have to be able to scale. It’s a key area of our attention.”
Enterprise executives must also give thought to how subnets interoperate. Users need to be able to wander around the building or campus seamlessly, without continually logging on or off. Permission access levels, security information, access to databases such as e-mail, and other applications must move with them. The CEO moving between areas will retain access to all department servers at all times, while a marketing employee, for instance, should only be able to access marketing data. The transition between access points must be handled with the utmost attention to security. As Clark said, “You have to manage the handoff.”
Part of the design task is mathematical—figuring out how many users are splitting a certain amount of bandwidth over a certain distance. The other element is the actual placement of the access points. “You have to physically locate the devices, and get a wired connection to them,” said Clark.
Once it's built, the “radio infrastructure” must be actively managed. “Initial deployments of access points may not produce the coverage required, and overlapping channels can cause interference," Clark said. "These issues must be discovered and corrected by adding more access points, reallocating channel assignments, or controlling power levels to address interfering access points.”
The procedure for figuring out where to place access points, according to Au, is fairly clear. First, the CIO and his staff should do a survey to see which frequencies are being used by WLANs on adjacent offices and on the floor above and below. The second step is to consult the information provided by the vendor on placement. It's a series of choices, depending on how much throughput a company wants at certain distances. Finally, temporary access points should be established at various points in the WLAN area. Collect data concerning the strength and purity of the signal, then make your decisions on permanent placement. Freeware tools are available for this process. AirMagnet offers a tool, Au said, that does this in a more automated fashion than the freeware tools.
CIOs must also recognize that users access their phones in more places—parking lots and rest rooms, for instance—than their computer devices. “Network coverage must be considered in this light,” said SpectraLink director of marketing Ben Guderian. Adequately supporting voice requires tight control of latency and jitter through a quality of service (QoS) standard. Currently, the issue is handled by proprietary algorithms. A standards-based approach to the problem is expected from the IEEE’s 802.11e committee sometime later this year.
WLAN voice has ramifications for both the wired and wireless network. Clearly, access points must be designed to maintain QoS, or the voice application will fail. However, the challenge goes deeper: Until now, most wired LANs have been purely data networks. “Thus, if the WLAN is intended to support voice—and most are—the LAN with which it will partner must be upgraded,” Guderian said.
Managing the standards risk
The ramifications are that the enterprise always has to be careful when facing a scenario in which there is a challenge today and proprietary solutions exist to solve it—but standards-based solutions are expected. The question is whether it's best to wait for the standards or use today's solutions if the desired application is deemed critical. Generally, if you go with a non-standard proprietary solution, you're running the risk that those vendors (or this particular solution) will disappear when the standards-based solution arrives. So, a few years down the road, a company that needs help with an antiquated proprietary solution may be out of luck.