Q: I recently read the LWAPP Flap article in Network World. The article suggests that the only currently existing alternative to using Lightweight Access Point Protocol (LWAPP) for communications between access points and wireless LAN switches is the GRE tunneling protocol. What are the impediments to mass adoption of GRE? – Josh, Chicago
A: GRE, or General Routing Encapsulated, is just a tunneling protocol. While it is a well accepted standard in pure wired applications, it was not designed with WLAN requirements in mind, and therefore has deficiencies.
There are two different types of traffic exchanged between lightweight access points and WLAN appliances — control and data. Data traffic is user data encapsulated in a tunnel and exchanged between the devices. Control traffic provides several functions, and is different between vendors. However, at a minimum, the control traffic provides functions such as tunnel setup, device configuration, rogue reporting, etc.
GRE, as defined in RFC 2784, provides a generic means of encapsulating traffic over an IP network. Therefore, using GRE, one could potentially encapsulate IPX, SNA and other protocols over an IP network. However, one of the challenges in the use of GRE is the establishment of the tunnel, which is typically done in a rather static fashion (other than PPTP, which defines a tunnel setup protocol and uses GRE to encapsulate PPTP frames). There is also currently no standard for encapsulating 802.11 frames within GRE, so forwarding 802.11 traffic within GRE requires either some proprietary Ethernet type, or that the frames be converted to 802.3.
LWAPP, on the other hand, is an open protocol* that addresses all the unique requirements of a wireless system. It defines both a control and a data component, where the data portion of the protocol specifically tunnels 802.11 frames between the access point and the appliance. The control component of the protocol provides such functions as firmware download (to update access point firmware), device configuration, user session management and tunnel setup. Furthermore, the control protocol is secured using public-key cryptography.
Any vendor currently using GRE has also created a proprietary protocol that exists in tangent with GRE. Although GRE in itself may be a standard, it does not provide most of the functions that are required for a next generation WLAN platform. To date, no one has introduced an open protocol in a GRE environment to any of the standards bodies to address these unique functions. Without this, GRE-based solutions will continue to remain proprietary.
With the adoption of an industry standard protocol for next generation WLAN devices like LWAPP, network administrators will be able to design, deploy, and redeploy wireless networks without fear of being trapped into one vendor due to proprietary solutions. Instead they will be able to base their purchasing decisions on the quality, features, and costs of available products.
* Editor’s Note: The LWAPP standard was originally drafted by Airespace and NTT DoCoMo, and is currently backed by vendors such as Chantry, Legra, Proxim, Symbol, Avaya and Intel. A final standard is expected in 18 to 24 months.
