MANHASSET, N.Y. A multivendor consortium, led by NTT DoCoMo and Airespace, has put before the Internet Engineering Task Force (IETF) a proposal for a standardized interface between wireless LAN access points and switches in centralized WLAN architectures.
If successful, the Lightweight Access Point Protocol (LWAPP) will fundamentally alter WLAN component design and deployment business models, while leveling the playing field for a plethora of centralized WLAN vendors vying for a toehold in a market dominated by Cisco Systems Inc.
LWAPP is intended to make the access points (APs) from the myriad vendors now deploying centralized architectures interoperable, thereby minimizing cost, increasing flexibility and lowering the threshold of WLAN entry to the enterprise.
"The proposal is the closest thing to a united front these startups have against Cisco," said Max Smetannikov, an analyst for networks and infrastructure at the451, which focuses on disruptive technologies. "I'm sure that anyone who's in the Wi-Fi space is watching this closely because it affects a lot of people," he said. "However, some are more active than others as they have products riding on it."
One of those companies is Chantry Networks (Boston, MA). "The bulk of the industry, both large scale enterprises and service providers, is trending toward a centralized architecture," said Brian Collie, co-founder and exec vice president of operations for Chantry. "We're looking at a $1 billion to $2 billion industry [worldwide] by 2007, so the standardization of this interface will have huge economic implications."
But before the impact comes the standards process, which is already mired in debate. For many vendors, the discussion of a protocol cannot even begin until the split of functions such as security between the AP and the central switch, or access router (AR), has been defined. This split directly impacts latency, a critical parameter for applications such as voice over Internet protocol (VoIP) and other time-sensitive applications.
As these functions are part of the media access control (MAC), the definition directly affects WLAN chip design, with obvious implications for the many companies now supplying those chipsets. As a result, the MAC-split issue raises the specter of charter overlap between the IETF and the IEEE 802.11 working group, upon who's work WLANs are based.
The parallelism between the nascent IETF work and the IEEE 802.11 working group is underscored by the IEEE's extensions to the standard (802.11k), which will define many of the radio-resource management functions that centralized-WLAN vendors are also attempting to define. At the recent 802.11k gathering in San Francisco, the IETF/IEEE liason, Dorothy Stanley presented the IETF's position in an effort to minimize any potential overlap.
Once the issues of distribution of functions between the AP and AR, as well as the politically charged issue of IETF/IEEE overlap are settled, work can then begin in earnest on LWAPP. Though many regard the initial proposal, authored by Airespace CTO Pat Calhoun, a good starting point, it is by no means a slam dunk. According to NTT DoCoMo's James Kempf, chair of the IETF's Control and Provisioning of Wireless Access Points (CAPWAP) group that is leading the effort, "Some elements will be included, but by no means will it be 'rubber stamped', the IETF doesn't work that way."
Issues to be addressed with the protocol, said the451's Smetannikov, have to do with a lack of emphasis on device discovery and subscriber management features. The former allows automatic configuration of newly added APs and impacts scalability. The latter means end-user information may be traveling back to the switch, unprotected.
"Airespace emphasizes AP management and wants the protocol adopted as is, which would give the standard a timeline of about 18 months", said Smetannikov. To push adopting, Airespace last week (July 22) said that it would not enforce a pending patent that it believes would be necessary to implement LWAPP.