After a few years of small-scale pilot tests, Wireless LANs (WLANs) based on the IEEE 802.11 standards are now moving into the mainstream for enterprise customers. This sudden shift into reality mode (i.e. "we have to buy something!") raises a number of perplexing questions as users develop specifications to guide that installation. The IEEE 802.11 committees and the Wi-Fi Alliance have developed a range of important new features, and those capabilities must be addressed in our design specifications.
The good news is that, overall, wireless LAN technology is maturing. The original crop of products was designed for the requirements of home users and small-scale commercial installations. These solutions could not easily grow to support networks incorporating dozens or hundreds of access points with thousands of users. We are now witnessing the introduction of tools that will allow us to build and maintain those large-scale wireless LANs with features to address the security, performance, and manageability requirements of commercial users. That also means commercial buyers will have to be able to sift through the options and develop a solution that will provide solid foundation from which to grow.
The purpose of this series is to identify and review five of the major developments in wireless LANs and provide some general guidelines regarding product selection and the potential pitfalls that will line the path. We will assume a basic understanding of WLAN fundamentals as we then describe these planning steps:
- Planning for capacity, not just coverage
- Moving to 802.11a
- Assessing security enhancements: WPA, 802.11i
- Incorporating quality of service with 802.11e
- Planning for manageability: Switch to WLAN switches
Critical Step 1: Planning for capacity, not just coverage
Providing access to a service that does not live up to expectations might be as bad as providing no service at all. In planning capacity per user, you must begin with the usable capacity of a WLAN channel and an estimate of how many users will be sharing it. One 11-Mbps 802.11b wireless LAN channel provides a real throughput of about 5.5 Mbps after we net out the impact of protocol headers, acknowledgements, retransmissions, and other network overhead. The 802.11a and g networks provide a maximum throughput of 30 Mbps, though that is reduced significantly if the 802.11g users are sharing a channel with 802.11b users. So the estimated throughput of a WLAN channel is roughly 50% to 55% of the raw data rate. That estimate assumes all users are in fairly close proximity to the access point and so are operating at the maximum data rate.
Proper design involves providing signal coverage in all areas, but also insuring that the network delivers adequate performance for all users in a cell. It should also be noted that wireless LAN design is an ongoing process, so these capacity issues must be revisited as usage grows and more access points/cells are added to the network. Of course, adding more access points also increases the cost of the network.
There are four basic factors in the design that we can control to insure adequate capacity as well as adequate radio coverage.
- Radio link interface: First we can choose the radio link interface that will be used, 802.11a, b, or g.
- Cell layout/channel assignment: The designers then select the placement of access points, antennas, and radio repeaters to provide coverage and to limit interference.
- Power levels: With a limited number of available channels, the same channels must be reused in different parts of the facility. When a channel is reused, we must reduce the transmit power of the other access points using that channel to limit co-channel interference (i.e. the interference created by access points in different parts of the facility that are assigned the same channel). However, reducing transmit power also reduces the range, so more access points may be required to provide the same coverage.
- Limit association rates: One last technique to keep low-speed users from impacting the overall performance is to limit the range of rates at which users will be allowed to associate. For example, in an 802.11b installation, you can limit association rates to users whose signal strength will support data rates ¡Ý5.5 Mbps. Higher transmission rates require a stronger received signal, so supporting only the higher data rates will mean more cells have to be provided or there will be dead spots in the coverage area.
|IEEE 802.11 radio link interfaces|
|Standard||Maximum bit rate||Fallback rates||Channels provided||Band||Radio technique|
|802.11||2 Mbps||1 Mbps||3||2.4 GHz ISM||FHSS or DSSS|
|802.11b||11 Mbps||5.5 Mbps
|3||2.4 GHz ISM||DSSS|
|802.11a||54 Mbps||48 Mbps
|12||5 GHz U-NII||OFDM|
|802.11g||54 Mbps||Same as 802.11a, plus 2 Mbps||3||2.4 GHz ISM||OFDM|
Go on to part two in the series, Moving to 802.11a.
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.