Keep track of the 802.11 alphabet soup
Which 802.11 standards will actually make a difference?
By John Cox, Network World | Network World US | Published: 17:00, 15 December 2004
The bad news is the alphabet of IEEE standards for wireless LANs keeps expanding. The good news is you can ignore a lot of it, at least until the new technology starts showing up in actual WLAN products.
Some projects in the 802.11 Working Group, which oversees WLAN standards, are nearing completion. Other groups, such as those working on mesh networks and for fast roaming, are just starting up.
Most network executives will never have to worry about the fine print in these technical standards, which cover things such as modulation schemes, access protocols and authentication, or (in wired networks) whether to transmit power in the Layer 1, or physical layer.
Most of the worrying is done by wireless chip makers, WLAN product manufacturers and software engineers.
Standards compliance can be meaningless
Of course the standards are only one part of the decision making involved in deploying an enterprise WLAN. Corporate suppliers might swear they support 802.11i, the recently completed standard that fixes many weak points in WLAN security. "But that's almost a meaningless statement," says Sheung Li, product line manager and 802.11 liaison for wireless chip maker Atheros. "The question is, what is the vendor actually offering in terms of capabilities and tools for a full security implementation?"
After all, standards rarely introduce a new technology. Instead, they create common ways for a WLAN to be created, monitored and managed. Vendors use the standards as a foundation and add unique or at least distinguishing features and functions atop the standard. In many cases, vendors will have their own code to do a given WLAN function and then replace it when an 802.11 standard is ratified. For example, Airespace has written its own code for radio resource measurement and management in its WLAN switch product line. When the 802.11k standard is ratified, the vendor will upgrade its products with code based on that.
Most vendors have at least one person who tracks or participates in these IEEE groups. They should be able to give customers a road map of their implementation plans.
The Wi-Fi Alliance, an industry trade group promoting 802.11 WLANs, is taking on an expanding role in creating certification tests for WLAN products as new IEEE standards are ratified. The alliance is finalising test programs for the 802.11d and h standards and has plans to introduce others next year (read our interview with Wi-Fi Alliance managing director, Franz Hanzlik) .
The Alliance keeps things above board
The Alliance also has a role in refusing to certify new WLAN features that vendors introduce ahead of final IEEE ratification. This work - surprisingly - has been controversial. Ira Brodsky of Datacomm Research has argued that refusing certification to non-standard products stifles innovation while others have pointed out (correctly in our view) that certifying non-standard products would be nonsense.
A few vendors, mainly in the consumer sector, have just started introducing WLAN access points that use intelligent antennas and special algorithms to boost throughput. This technology, called multiple input multiple output (read our explanation of MIMO), is a likely candidate for the next faster speed WLAN standard, 802.11n. But that work has only just begun.
"We want to discourage any use of terms suggesting that (802.11n) is real and that products can be 'compliant'" at this stage, says Frank Hanzlik, the alliance's managing director.
802.11n and the WLAN standard projects
The 802.11n Task Group definitely has the starring role in 802.11 standards work. It is charged with creating a standard for WLANs with at least 100 Mbit/s throughput, compared with 20 to 25 Mbit/s available today. The group is sifting through scores of technical proposals. Don't expect a final standard until sometime in 2007 (and read our latest update on 802.11n progress).
But there are plenty of other WLAN advances completed or in the works. Some of the lesser-known or more recent standards work on 802.11 are:
- 802.11d, multi-country roaming: Old by WLAN standards, the working group finalized 802.11d in 2001. It's a way for WLAN access points to broadcast what country they're in and what country-specific rules client network interface cards have to follow. You can fly from New York to Rome, walk into your office or hotel, fire up your wireless laptop and expect to connect with whatever WLAN is open.
- 802.11F, inter-access point protocol: The capital "F" designates a "recommended practice" not a formal standard in IEEE parlance. The basic idea was to create a way for access points to talk among themselves and to transfer data associated with a connection quickly from one access point to another. But WLAN vendors already have figured out how to do this on their own for their own access points. The document was published in mid-2003, but so far is not being embraced. "There are a set of technical issues that make this (hand-off) difficult," says Andrew Myles, manager of wireless standards for Cisco's Wireless Networking Business Unit. "As far as I know at this point, no one is planning to support it." (for more on wireless roaming, see WLAN hurdles 3 and WLAN roaming - the basics)
- 802.11h, dynamic frequency selection, and transmission power control: the original idea for 802.11h was to create a set of management messages for access points and clients in the European 5-GHz band to coordinate efforts to avoid interfering with radar and satellite communications in the same band, Atheros' Li says. The WLAN devices select another channel and adjust power output if needed. But these same actions can be used to improve WLAN efficiency, quite apart from any specific regulatory rules, according to Li, as more countries open more 5-GHz bandwidth for use by 802.11a WLANs. The standard was final in September 2003 (read a White Paper on 802.11h. Products with some of these features might start appearing soon. Some elements of this work are being carried into another standard, 802.11k.
- 802.11j, use of the 4.9- to 5-GHz spectrum in Japan: Originally, this group's work was focused on making changes to the 802.11 media access control and 802.11a PHY layers to gain Japanese regulatory approval in this band. But, with the standard newly completed, Li points out that the FCC recently allocated this same band for licensed spectrum set aside solely for public safety and homeland security in the US. The 802.11j work on how to use this spectrum could prove useful in the US as vendors introduce products for public safety networks in this band.
- 802.11k, radio resource management: Launched in late 2002, this project will standardise an array of radio measurements, dealing with roaming requests, an array of data about the radio channel and data about the client devices. In addition, this data can be made available to higher-level WLAN management applications, where the information can be used in tasks such as optimising performance and balancing traffic loads. Task Group K is sifting through some 1,000 comments that members submitted on the proposed standard during a recent ballot, says Clint Chaplin, wireless standards lead at Symbol Technologies. Completion is likely in mid-2005. According to Cisco's Myles, the Wi-Fi Alliance will launch a testing and certification program for 802.11d and k next year.
- 802.11r, fast roaming: Handing off clients quickly from one access point to another with their authentication and security policies intact becomes critical when clients are moving, such as VoIP calls made with handheld WLAN phones, Chaplin says. This group, launched in 2004, is creating a standard way to make roaming fast, so that users don't have to re-authenticate at each new access point or have their calls disrupted, he says.(see WLAN hurdles 3)
- 802.11s, wireless mesh for access points: Formed in early 2004, this group is creating a standard that will let access points act as routers for wireless data, forwarding traffic to neighbouring access points as Internet nodes do today with a series of multi-hop transmissions. Such mesh networks are inherently more reliable because they can route around failed nodes, and can adjust to balance traffic loads and optimize performance. The members still are sorting out their approach and schedule. Symbol's Chaplin says the first call for proposals is likely to be in January.