802.11a can significantly expand the power and flexibility of 802.11b wireless LAN deployments, but in most cases, this power is being squandered. Typically, organizations using traditional 802.11b WLAN technology are being advised by their vendors to overlay an 802.11a network to fill in coverage gaps and extend network coverage. But what if a network engineer could obtain pervasive WLAN coverage with 802.11b alone? In that case, 802.11a could be used to support priority traffic, high-bandwidth applications, or growth in the user base. This article discusses the challenges of achieving pervasive coverage with 802.11b, and suggests a new WLAN architecture that overcomes those challenges.

THE CHALLENGE OF PERVASIVE COVERAGE

As companies have noted the impact of wireless access on employee productivity, they have begun to see the need for more pervasive WLAN coverage beyond hotspots in conference rooms. And with the arrival of more advanced applications such as mobile voice over IP (VoIP), pervasive coverage becomes essential to the proper function of those applications. With the largest penetration of client devices, 802.11b is the likely choice as a platform for such pervasive coverage. But whether a company uses 802.11a, b, or g, there’s a fundamental problem with using traditional access points (APs) to provide pervasive coverage.

The 802.11 protocol and the chipsets used for 802.11 APs were designed to support wireless LAN hotspots consisting of one AP. As companies extend coverage by deploying adjacent APs, co-channel interference rears its ugly head.

Co-channel interference results from two APs transmitting on the same channel in the same area. In a typical 802.11 wireless network, APs continually broadcast a signal, and any wireless client within range can lock onto that signal to make a connection. But when APs are placed close enough together to provide continuous coverage over a large area, it’s possible (or even likely) that some APs will generate signals that interfere with other APs. When this happens, two or more APs transmit packets at the same time on the same channel, corrupting packets and causing clients to experience transmission delays and lower performance. In addition, clients located equidistant between two APs on the same channel may flip-flop, not knowing whether to access one or the other.

To minimize co-channel interference, most WLAN vendors recommend that users operate adjacent APs on alternating channels. 802.11b and 802.11g have three available channels (1, 6, and 11 in 802.11b, for example), while 802.11a has 11 available channels. Enterprise WLAN vendors offer site survey and management tools to help users plan AP deployments with alternating channels to maximize coverage while minimizing the possibility of co-channel interference. These systems may also allow users to regulate the power output of individual APs to make their coverage areas larger or smaller, and so fit better into an overall coverage map.

Unfortunately, these measures do not eliminate co-channel interference. With only three channel possibilities available in 802.11b or g, it’s inevitable that companies deploying large WLANs will either face co-channel interference or they’ll have to establish coverage gaps to prevent it. This is why traditional vendors look to 802.11a’s non-interfering frequency and 11 channels as a band-aid to help avoid co-channel interference. These vendors recommend overlaying an 802.11a network and then assigning certain groups of employees or applications like VoIP to the new WLAN. But there are two key drawbacks to deploying an overlay WLAN: cost and complexity.

Cost--With two separate WLANs to manage, companies must spend twice as much to manage their WLAN infrastructure. In addition, users assigned to the 802.11a WLAN must be given 802.11a client cards at extra cost rather than using the 802.11b or b/g cards that may be built into their laptops and PDAs.

Complexity--As furniture is moved, walls are reconfigured, or user densities inevitably change in the WLAN coverage area, the challenge of reconfiguring APs to match the new space becomes doubly difficult when there are two WLANs to deal with. In addition, using 802.11a’s 11 channels may be a way to avoid co-channel interference but in using that band-aid one makes it exponentially more difficult to adjust for changes in the environment than the 3-channel 802.11b/g technology.

As WLANs move out of hotspot applications and become pervasive extensions of corporate wired LANs, the world will need simpler and more permanent solutions to the co-channel interference problem.

CELLULAR NETWORKS VS. WIRELESS LANs

Cellular phone networks provide an excellent model for the ideal WLAN architecture. Unlike WLANs based on APs designed for standalone use, cellular networks were designed from the ground up with many base stations supporting pervasive coverage for large numbers of users with high quality of service. When we compare WLANs based on stand-alone APs with cellular network architecture, there are two key differences: network-controlled access and AP coordination.

Network-controlled access

In a typical WLAN, the clients control the transmissions: the APs broadcast signals and accept requests for connections from any client. When several clients are closest to a particular AP, they compete for access on a first-come, first-served basis. Since the AP is a passive partner in this arrangement, it’s impossible to optimize individual client access or transmission characteristics (offering different service levels to voice users as opposed to e-mail users, for example). In addition, clients may choose to connect to the nearest AP whether or not that AP is already overloaded.

But in a cellular network, the network individually controls each client’s access to an AP. Because the network is in control, it can optimize each client connection for that client’s specific performance and QoS requirements. The network can also minimize denied accesses and optimize performance by load-balancing client traffic among APs.

AP coordination

WLANs do not have coordination among APs, while cellular networks do. In a cellular network, every base station knows about every other base station, and the network manages base station transmissions to avoid co-channel interference and create a continuous blanket of coverage. This feature gives cellular network architects the flexibility to deploy new base stations wherever they are needed without having to alternate channels. In fact, cellular network architects typically overlap base station coverage on the same channel to provide continuous coverage.

When it comes to wireless LANs, then, an architecture that offers network-controlled access and AP coordination offers a far less complex, less costly, and more permanent solution lies. This approach offers several important benefits for 802.11b performance and deployment simplicity, and it allows 802.11a to be deployed as needed for power users or complex applications.

A CELLULAR WIRELESS LAN ARCHITECTURE

Since the basic problem of traditional wireless LANs begins with APs designed for stand-alone use, the solution begins with APs that have been designed for coordinated use. Rather than using commodity technology that can’t coordinate or control client access, the new WLAN architecture uses new technology designed to enable a cellular WLAN architecture while remaining fully compliant with 802.11 standards.

The new cellular WLAN architecture eliminates the problems caused by co-channel interference, delivers pervasive WLAN coverage with 802.11b, and enables the use of 802.11a for adding power or capacity to the basic infrastructure. In the process, the WLAN becomes a much more transparent extension of the wired LAN from a management and engineering point of view.

Eliminating co-channel interference problems--To eliminate co-channel interference issues, all APs in the new WLAN architecture coordinate, timing their transmissions to avoid interfering with one another. This allows multiple APs to transmit on the same channel without losing bandwidth and performance to co-channel interference. In addition, the client transmissions also are timed to avoid interfering with other transmissions.

Simplifying deployments--Since AP coverage areas can overlap without fear of co-channel interference, there are no coverage gaps. Network engineers can simply place APs to deliver strong signals throughout the coverage area. Moreover, network engineers do not have to pay for RF analyses or site surveys, because there are no channel assignments or AP power levels to manage.

Improving quality of service--Because the new WLAN assigns each client to a specific AP, the client doesn’t flip-flop between APs. APs can recognize and prioritize client access based on application or user level to ensure guaranteed quality of service for VoIP or other important traffic. The new WLAN also avoids the simultaneous requests for access to any AP (a primary cause of performance problems), because it enables optimal load balancing among APs so none of them becomes overloaded. Once a client is connected to an AP, the client’s connection quality doesn’t vary as other users request access or are handed off to another AP.

Speeding client hand-offs--With a cellular WLAN architecture, the wireless network looks like one giant AP to the client, even though each AP has its own MAC and manages full line-rate bandwidth. The system eliminates the handoffs that occur when the client roams from one AP coverage area to another, because the client never sees a change in AP. (Such handoffs take up to two seconds with traditional WLANs). Seamless handoffs also improve service quality for VoIP and other delay-sensitive applications.

Making the WLAN more manageable--Solving these problems creates a fundamental change in the relationship between the network manager and the wireless LAN. Since network managers no longer have to deal with wireless frequency issues (co-channel interference and AP channel assignments), they can treat the WLAN infrastructure as a seamless part of the overall LAN system.

ADDING POWER WITH 802.11a

This new WLAN architecture also allows network managers to use 802.11a as a power enhancement, rather than as a coverage solution. With a pervasive 802.11b WLAN in place, network managers can use 802.11a to provide high-performance access for priority users or applications, or to add bandwidth to the overall WLAN as it becomes necessary.

Since enterprise WLAN rollouts began, co-channel interference has been the dirty little secret of traditional access points. Vendors of such access points would have their customers believe that complex AP deployment maps, poor performance in dense user populations, and 802.11a overlays are the only way to solve the problem. But optimized WLAN service doesn’t have to be so complex, and 802.11b can perform far better than we have been led to believe. A cellular WLAN can be as simple to plan and deploy as a wired LAN, can deliver pervasive 802.11b coverage, and can leverage higher speed technologies like 802.11a for additional performance.

Joel Vincent is Director of Product Marketing for Meru Networks.

Visit Meru networks online at www.merunetworks.com.