Perhaps the Enterprise Is No Place for 40MHz

Meru Networks recently announced its 802.11n equipment, becoming the first enterprise vendor to go public with plans for the new wireless specification. Meru has some natural advantages when it comes to 802.11n, but I am starting to question how much impact some of these advantages will actually have on enterprise

Meru Networks recently announced its 802.11n equipment, becoming the first enterprise vendor to go public with plans for the new wireless specification. Meru has some natural advantages when it comes to 802.11n, but I am starting to question how much impact some of these advantages will actually have on enterprise adopters.

Meru utilizes a single-channel architecture, which, in a nutshell, means that all access points in a WLAN broadcast on the same channel and the back-end wireless controller handles all the interference and handoff-related issues. The single-channel architecture will effectively let them advertise a single 40MHz channel in the 2.4GHz band, something other vendors will struggle to match.

Most other wireless vendors need their APs to be on separate channels to avoid interference. Since there are only three nonoverlapping channels in the 2.4GHz spectrum, administrators need to set things up pretty carefully to minimize the impact of two nearby APs broadcasting on the same channel—taking care to account for the impact of their own network and neighboring networks as well.

Well, 40MHz channels really screw up that model. Instead of three nonoverlapping channels in the 2.4 band, with 40MHz channels there is now only one. Doing the math, it is pretty hard to have nonoverlapping channels on neighboring APs when there is only one channel.

The IEEE 802.11n task group recognized this problem awhile ago and, in the last draft, instituted the recommendation that 802.11n transmissions in the 2.4 band only use 20MHz channels by default, and that 40MHz be more of a 5GHz band thing.

So Meru can do something others may not be able to do, but really, how relevant will this advantage be?

40MHz channels could definitely provide more throughput headroom than a 20MHz channel. The link rate will be higher (300M bps versus 145M bps), and the resulting real throughput will also be similarly increased. But at either data rate, 802.11n will provide much more throughput than 802.11g—whether it is 3x or 6x.

But the channel management dynamic will already be changing. Via multiple antennae, and MIMO, 802.11n access points will automatically get better range performance than 802.11g APs. So access points that did not interfere with each other in the past will now cause problems with each other because of the improved range. Administrators will need to retune their networks, dialing down transmit power or rethinking the size of their cells and the access point placement. But the new cells still need to support a fleet of legacy clients that do not yet support the new standard, so there is only so much rejiggering that can be done without interrupting legacy client service.

And one thing is for sure, enterprise wireless assessment tools are lagging behind the access products, so administrators will not have access to software that will adequately tell them what they need to know to plan out the new network and all it entails. And 802.11n will require a lot of planning before deployment—on both sides of the network.

To reduce the question marks that will come with 802.11n adoption out of the gate, I foresee two different courses of deployment. Wireless administrators will either a) layer on a new wireless network that is 802.11n-only in the 5GHz band, while completely shutting off legacy 802.11a networks, or b) try to build out 802.11n within existing networks and rely on the central controllers to advise and guide the migration.

If the former, the administrator could likely implement 40MHz channels. If the latter, however, the complications added by the wider channels will only confuse the deployment, without adding enough benefit. In that case, I foresee the use of 20MHz channels only until some tipping point is reached in 802.11n client penetration.

Either way, I don't foresee the use of 40MHz channels in the 2.4GHz band as any kind of benefit for the next several years, so Meru's architecture advantages may be little more than a pretty marketing presentation in this case.