* 802.11n could find near-term niche in backhaul nets
As 802.11n grinds through the IEEE standards process, enterprises eagerly anticipate its arrival. But they are understandably reluctant to invest heavily in today’s pre-standard products.
In fact, just 1% of about 300 wireless LAN professionals surveyed recently by Kubernan indicated near-term plans for 802.11n (see the full “Kubernan 2007 Wireless LAN State-of-the-Market” report here).
There is one enterprise application for 802.11n, however, that might be considered safe for the near term: using 802.11n access points in a mesh configuration for wireless backhaul.
In such deployments, the “n” devices sit in the core of the wireless network. They receive and aggregate traffic from 802.11g or other lower-speed access points and pass them over a wireless 802.11n hop or two to an edge access point wired to an Ethernet switch. Mesh saves time and money when new cabling is required to further build out the Wi-Fi infrastructure; in traditional models, each access point must be cabled to the wired LAN. Indoor mesh is also useful in difficult-to-cable areas or where asbestos suddenly rears its head and an air link is needed to circumvent physical cabling.
There are a couple of attractive characteristics for even pre-standard 802.11n in this role: The speeds (usually up to 300Mbps) are better suited for aggregating 802.11g traffic, which runs at a maximum of 54Mbps. It is more efficient for lower-speed networks to aggregate up onto higher-speed backbones.
Also, as with many existing outdoor mesh networks, the 5GHz band could be used for the 802.11n backhaul and the 2.4GHz band could be used for 802.11g (or legacy 11b) client access to ease channel planning.
Finally, you likely only need to invest in a handful or so of products – unless you are a very large, distributed enterprise entirely throwing out your cabled Ethernet network. The 802.11g access points can stay where they are, and you can avoid new cabling by using 802.11n. You get some experience with 802.11n and hopefully improve overall throughput. If – heaven forefend – the pre-standard products wind up incompatible with the final standard, you haven’t invested in hundreds or thousands of them.
What about vendor support for this model? Find out next week!
