A challenge since the dawn of the wireless local-area network (WLAN) era has been to provide adequate coverage. This becomes even more pressing as voice and other interactive applications are added. The additional complexity is because people use their phones in stairwells, hallways and other areas where they don't often carry out data-only tasks. In addition, voice lowers tolerances and makes delivery of near-pristine signals more important.
This feature by consultant Jim Geier does a good job of describing interference problems and providing some ways to deal with the issues. Geier says that the 802.11 protocol was designed in such a way that one station will back off if another is sending a packet. The problem is that interference can look like legitimate traffic, causing other stations to hold their fire. Geier points out that the situation is exacerbated by the fact that the interference itself follows no such etiquette and transmits in an unpredictable manner.
There are several types of RF interference, all of which are outlined in this post. The writer says that there is physical interference, narrowband interference, wideband interference, inter-symbol interference and, in a category that clearly doesn't sound like very good news, all-band interference. A definition of each of these is offered.
Geier offers five ways to limit problems: Analyze the possibilities for interference before deployment; shut down the source; provide strong enough signals to overcome some or all of the interference; configure systems properly; and opt for 5 GHz deployments.
The air in enterprises clearly is getting more crowded with an increasing density of different types of wireless signals. The newest entrant is 802.11n, which is still in pre-standard form. At the Interop conference in Las Vegas, representatives from five companies -- Xirrus, Aruba, Siemens, Aerohive and Extricom -- participated in a panel discussing WLAN architectures. While not a direct transcript, this eWeek post is a seemingly comprehensive blow-by-blow account of the presentation. Topics discussed included: misconceptions on 802.11n; which architecture is best; how to handle the cabled side of the equation; the best method of adding services; and what to look for in a client.
A deeply related issue to RF interference is the way in which access points (APs) in a WLAN communicate with each other and with the central hub, which also can house an AP. The predominant method to date is one in which all the APs link to the hub, which is wired to the host network.
An alternative approach -- mesh networking -- is gaining adherents. Rob Smith, the secure networks group chief for network archtecture for Telos, told me in a recent interview that a mesh is a configuration in which APs connect with one or more of their cousin APs instead of individually reaching out to a central hub.
This approach can be cheaper, more flexible and faster to set up, since wiring and powering requirements are reduced. This ZDNet.co.uk piece describes Ruckus Wireless' SmartMesh technology. This system, which was introduced last month, is designed for businesses.