Wi-Fi vendors duke it out over airtime fairness

* Managing client access vs. optimizing client performance

Last week I discussed airtime fairness in one newsletter and beamforming in the other. Airtime fairness and beamforming are separate technologies (usually, anyway). But they do have something in common: They both help control transmissions in ways intended to give Wi-Fi users better and more predictable wireless experiences.

However, as it turns out, Cisco considers beamforming and airtime fairness to be more or less synonymous within Cisco Wi-Fi networks. Unlike its competitors, Cisco attacks the airtime fairness issue with its recently announced beamforming technology (called ClientLink). ClientLink better focuses RF energy toward legacy 802.11a/g clients to improve their performance - whether there are any 802.11n clients in the network or not.

However, one side effect is that if the performance of 802.11a and g clients in a mixed 802.11a/g/n network improves, that will free up more airtime for the 802.11n client to transmit, explains Chris Kozup, senior manager of Cisco’s mobility solutions.

“The other vendors limit legacy clients to make sure 11n devices are having their say,” Kozup asserts.

Those other vendors beg to differ.

“The concept of airtime fairness is well known,” says Joe Epstein, chief architect at Meru Networks, an airtime fairness pioneer. Meru, Aruba, Motorola and, most recently, Aerohive, for example, basically attempt to prevent the slowest client on the network at a given time from bogging down the rest of the Wi-Fi network.

To do so, they dynamically determine the exact amount of airtime each client is consuming in microseconds. They then adjust the number of opportunities each client gets to transmit using algorithms that account for each client’s characteristics, such as current throughput, distance from the AP and even 802.11n’s comparative efficiencies such as packet aggregation and 40MHz channels. The primary goal is to ensure that the slowest client doesn’t set the pace for all clients on the network and monopolize the air.

Epstein believes “it’s a stretch” to describe beamforming as an airtime fairness feature.

For its part, when Aerohive announced its Dynamic Access Scheduling earlier this month, it asserted that Meru, Aruba and others’ airtime fairness algorithms use protocol-based scheduling. Adam Conway, director of product management, said the competition makes assumptions about the client’s transmission speed based only on its maximum connect rate as defined in the 802.11a, b, g, or n protocol, not accounting dynamically for throughput, distance and other factors.

Not true! counter Meru and Aruba, who both say they use dynamic time measurements of client transmissions to divide up the airtime fairly.

From there, different vendors have different mechanisms to optimize each given client’s performance by assigning weights based on different variables, such as application, client type, SSID, user and so forth. This way, certain clients can gain preferential treatment over others where appropriate.

“Airtime fairness is necessary but not sufficient for best network performance,” says Mike Tennefoss, head of strategic marketing at Aruba. Aruba added airtime fairness last summer to a bandwidth-based fairness technique it already supported.

Indeed, airtime fairness is just one tool in the RF management toolbox.

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.

Copyright © 2009 IDG Communications, Inc.