Step 1: Planning for capacity, not just coverage
Step 2: Moving to 802.11a
Step 3: Assessing security enhancements
Step 4: Incorporating quality of service
Critical Step 5: Manageability -- Switching to WLAN switches
One safe assumption in a data network is that traffic will always expand to fill the available capacity-- and three times faster than you thought it would! That means that we must be prepared to accommodate growth and expansion, which will mean adding more access points. Each access point must be assigned a radio channel, and the WLAN architecture begins takes on the appearance if a cellular telephone network. However, we have a limited number of channels we can assign (three in the 2.4-GHz band and 12 in the 5-GHz band) and as we reuse channels in other parts of the coverage area we must take pains to limit co-channel interference.
Up until now, insuring adequate coverage and network capacity has involved a process of trial and error. WLAN site planning involved selecting "best guess" locations for access points, powering them up, and then wandering around with a test set to measure signal power and potential data rates in each part of the coverage area. As new access points are added, each must be assigned a channel, and the transmit power of other access points using that same channel must be reduced to limit co-channel interference.
Of course, if you reduce it too much, you wind up with "dead spots" or areas where there is no coverage at all. The result is an ongoing process of trial-and-error to position access points, select channels, and tweak transmit levels. Hopefully someone will be updating the records as we work through these reconfigurations or finding those access points again will become the great Easter egg hunt!
This is the area where the impact of wireless LAN switches will be greatest. For those unfamiliar with the concept, a wireless LAN switch describes a configuration where the functions of a number of specially designed "thin" access points are coordinated through a central server. While the level of sophistication varies from product to product, wireless LAN switches will typically incorporate a mechanism for managing the radio domain. They usually come with tools that allow them to insure adequate radio coverage, identify problem areas, and facilitate network upgrades.
In a wireless LAN switch environment, when a new access point is added to the network, the transmit power of the surrounding access points is automatically adjusted to reduce interference and maximize performance. Further, as the central controller will know all of its access points, it can quickly identify "rogue" access point installed by users who didn't read our security directives and "spoofed" access points installed by hackers seeking to gain access to our wireless LAN.
Besides managing the radio domain, wireless LAN switches can also centralize security management and record keeping and provide a solution that is geared toward large-scale commercial implementations.
The big question here is: Who will survive the inevitable shakeout? There are wireless LAN switch products available from well over a dozen vendors, including Airespace, Aruba, Chantry Networks, Reefedge, Trapeze Networks, and Symbol Technologies. What they all have in common is the need to do battle with Cisco. Cisco is moving into the area cautiously with their Structured Wireless Aware Network (SWAN) strategy, but given the dominance of their Catalyst product line in the wired LAN space, Cisco will have the inside track on most WLAN switch installations.
Again, we will face the trade-off between immediate functionality versus a standards-based solution. A mix-and-match strategy where a user could buy access points and central WLAN controllers from different vendors will require a standard for communicating between the central controller and the thin access points. Currently, each WLAN switch implementation is proprietary. However, there are two development efforts called the Light Weight Access Point Protocol (LWAPP) and the Architecture for Control and Provisioning of Wireless Access Points (CAPWAP) that seek to provide multivendor interoperability.
But standards are still a long way off, and we will have to wait and see if any can deliver the full suite of capabilities we get today in a single vendor, proprietary implementation. Most analysts are anticipating proprietary WLAN switch implementations for the next three years at least.
Things to remember
As you can see, there is no "one-size-fits-all" plan for wireless LANs. Commercial users who are looking to deploy these networks in an enterprise environment must recognize the evolving nature of the technology and decide how these new capabilities should be addressed in their network plans.
At the most basic level, planning for a wireless LAN involve four major questions:
- Who will be provided access (it is important to address the requirements of visitors as well as employees)?
- What level of performance will be provided to the different classes of users?
- Where will the service be available?
- How will we ensure that we are able to manage and maintain the operational and security issues introduced by this new network resource?
One configuration we are starting to see is the dual-overlay network. A low-capacity 802.11b/g network with minimum security and access privileges is deployed to support simple Internet access for visitors, while a higher capacity 802.11a network is provided for employees. To reduce equipment requirements, it is important to locate access points that are able to support both 2.4-GHz and 5-GHz channels and to support multiple WLANs from the same unit. As the visitors' network will be providing thinner coverage, you will probably require fewer 802.11b/g access points.
One questionable marketing tactic to be aware of is the WLAN Benefits Calculator. The calculator is an Excel spreadsheet program developed by the Wi-Fi Alliance with the help of the Gartner Group that allows you to compute the savings generated through the implementation of a wireless LAN.
As a general rule, you should be very skeptical about tools that assign real dollar values to time savings (e.g. each user saves x minutes per day, and if the average wage rate is y, basic multiplication will show that spending this much on a wireless LAN will actually saves you that much). Assuming that everyone will make good use of that additional 30 minutes of productive time requires a major leap of faith. In short, no one is sending you a certified check for the "savings." Nevertheless, if you'd like a copy, it is available at https://www.wi-fi.org/opensection/ wlan_calculator.asp. Just make sure your boss buys into the logic before you stake your proposal (and your career prospects) on the results.
Michael Finneran is an independent telecommunications consultant specializing in wireless networks and technologies. Besides his research and consulting activities, he writes a regular column called "Network Intelligence" for and teaches their seminars on wireless technologies and wireless LANs. He can be reached at firstname.lastname@example.org.