Skip Links

Aerohive Tackles the Scheduling Problem

By Craig Mathias on Fri, 02/06/09 - 9:10am.

Aerohive this week announced an important addition to their product line that continues with a theme we've seen from Aruba, Cisco, and Meru in recent weeks, to wit - that traffic scheduling is now a more important approach to improving performance than simply integrating the latest radio chip. Indeed, there'll be gigabit wireless LANs involving new radios within the next two or three years, but 802.11n is likely to be the mainstream technology for most of us for the foreseeable future. While I do expect continual improvements in radio performance, these won't yield the dramatic throughput improvements we've seen throughout the history of WLANs. So the best place to look for performance benefits is in how traffic is managed once it's on this side of the AP.

Aerohive's approach here is a bit different from what their competition has done, which is as expected given their distributed-control system architecture. But the concept in general is the same - schedule how data is moved over the air so as to optimize for the best mix of throughput by Wi-Fi technology. The problem in the case of mixed-mode (say, b/g or b/g/n, all on the same channel) deployments can be severe, akin to Porsches stuck behind, well, a slower, older vehicle going well under the speed limit in the fast lane. Queuing theory can be applied here, just as it has been in operating systems, disk drives, and waiting lines at the bank, to optimize service based on traffic mix. This isn't an exact science, and results can be further influenced by internal elements like group policy settings, traffic prioritization based on QoS requirements, and externalities like instantaneous interference. But Aerohive's announcement of their solution here is a powerful indication that scheduling theory has become just as important as great Wi-Fi chips in achieving the best possible WLAN performance. The company provided both freespace and lab (using Veriwave's WiMix) measurements showing that a real benefit is present.

Aerohive also announced, BTW, an entry-level embedded management system for small networks, configurations that are often relegated to unmanaged SOHO- and residential-class products. It's often the case that small WLANs grow into big WLANs, and being able to do that without a wholesale replacement of infrastructure is pretty cool. I think this approach will be adopted by other vendors, and Aerohive is to be commended for taking the lead here.