* Mechanisms for deriving better and more predictable Wi-Fi performance
The IEEE ratified the bare bones Wi-Fi standard in 1997. Since then, equipment vendors and various 802.11 working groups have been beating it into submission with formal extensions and proprietary capabilities to render it a business-grade network. Much of the ongoing work relates to controlling the shared airwaves in ways that prevent big transmissions from any one client from squeezing out other clients’ traffic.
Basic 802.11e QoS standards are in place and implemented in products for packet prioritization and call admission control (CAC), a congestion avoidance mechanism that applies to voice traffic. Still, given Wi-Fi’s inherent shared, contention-oriented media access control (MAC) mechanism, getting reliable performance across the network requires more specialization.
One issue is that the amount of airtime consumed by a packet depends on its data rate. The Wi-Fi system needs to take that into account when scheduling transmissions. In other words, the system should allocate more opportunities to transmit to faster clients so that they get their fair share of the airtime. Otherwise, they can be bogged down by older and slower clients.
To address this issue, many of the Wi-Fi vendors have add-on QoS mechanisms of their own.
Startup Aerohive is the latest with an enhancement in this area. Its Dynamic Airtime Scheduling is somewhat analogous to Weighted Round Robin QoS capabilities in IP routing, in that it allows clients to take turns transmitting, but schedules the transmissions weighted toward variables specified by policies you set.
For example, clients far away from an AP transmit at slower speeds than those close up and, as such, consume more airtime. Dynamic Airtime Scheduling accounts for that. Once a packet is sent to the head of the transmission queue from a given client, that client’s airtime gets debited, explains Adam Conway, Aerohive director of product management.
A number of other vendors, including Aruba Networks and Meru Networks, have similar overlay mechanisms for divvying up the airwaves fairly. From these vendors, the capability is often called “airtime fairness.”
Niall Pariag, senior network administrator at Riverside Health, a 450-bed hospital in Yonkers, N.Y., is using the Aerohive solution. In a final vendor-selection bake-off, Pariag had compared Aerohive and Meru Networks’ airtime scheduling capabilities.
According to Pariag, Meru’s mechanism “assumes that an 802.11n client is faster than an 802.11g client.” But in reality, an 802.11n client can actually transmit at a slower rate than an 802.11g client, depending on its distance from the AP. “So Meru’s solution will only work some of the time,” he says.
