Tuesday, February 26, 2008

The four generations of wireless LAN networking

1st generation: early consumer Wi-Fi equipment, with minimal amounts of security

2nd generation: first release of enterprise-oriented wireless products. Access points still stand alone, but security was improved.

3rd generation: most current enterprise wireless products. Central controllers help regulate AP connections, much improved security, but deployment issues remain: proper layout and channel spacing needed for maximum benefit.

4th generation: All access points share a channel, with a central controller determining which access points communicate with various devices. Denser implementations become possible without risk of co-channel interference.

Fourth-generation APs, currently developed by Meru and Extricom, use a smart, centralized controller to create a large, virtual wireless cell that spans several APs, making the handoff between cells transparent to endpoint devices such as laptops and, ideally, reducing dropped connections as a user moves around a wireless LAN. As far as the device can tell, an entire office is just one large wireless zone (see sidebar for a look at the other wireless generations).

Extricom's APs have much less intelligence than typical offerings. They act similarly to antenna extensions that are intelligently tuned in to the appropriate device by a central switch to which each AP is directly connected. Meru's devices, on the other hand, have more intelligence in the AP, allowing them to communicate to the network on layer 3.

Each company claims its technology is superior in several ways, but King said they were similarly capable for most tasks.

These fourth generation methods also have the benefit of reducing planning complexity -- no more careful spacing of APs at set intervals, overlapping – but not by too much – connectivity zones to provide maximum range and throughput. Instead, these options can pack APs more closely to ensure stronger cover without the fear of radio interference.

But is it time to jump aboard? Maybe, maybe not. The technology is promising, but it is so new that network architects don't understand all its pitfalls.

See more info on wiki

0 comentarii: