With 2005 behind us, I wanted to take a minute to step back and reflect on the major wireless LAN stories from the past year. In my unscientific opinion based on what I remember, here are the top stories:
-
The biggest story in 2005 was the emergence of municipal wireless LANs. I started the year as a skeptic because of the inherent problems in building a reliable service on unlicensed spectrum. What I didn’t realize is that much of the communications equipment these networks are replacing are based on horribly slow networks. 802.11 may not be as reliable as a licensed network that operates without interference, but you can retransmit several times at 1 Mbps and still beat an ultra-reliable network that operates at kilobit per second speeds. There are a number of engineering challenges in building a large-scale 802.11 blanket, but cities have started to realize that their ownership of some public infrastructure (say, the poles for traffic lights) gives them a valuable asset in building these networks.
The battle is over, folks. Most people agree that it’s a good idea to be building these networks. The question that remains is how. The engineering is in its infancy, and a lot of people are going to have to learn a lot about how to do it very fast if growth will keep up.
-
The flexibility of wireless networks continued to be on display. Before the emergence of municipal networks, Main Street isn’t someplace that you’d expect to have a high-speed Internet blanket. Neither is public transit. In the past couple of years, however, a good number of transit systems are providing 802.11 access. (In addition to Connexion, see the Bainbridge Ferry, the Altamont Commuter Express (ACE) train, and my own sighting on the Tokyo Metro.) The use of 802.11 for the last hop obscures the hard part, which is that you need to somehow collect the bits from the 802.11 network and send them off to the Internet. Large-scale wireless area blankets are useful for this, whether satellite (as in Connexion) or some other vehicle-to-Internet technology (such as, say, a pre-WiMax mesh connecting a commuter train). Sadly, this trend does not yet include any of the transit systems I use, so I am still stuck with a relatively slow connection on most of my train trips.
-
Unlicensed spectrum is important precisely because there is no control over how it’s used. That didn’t stop Massport, the operator of Boston’s Logan Airport, from trying to use its powers to prevent Continental from setting up free service in their elite lounges. The freedom to set up a service using unlicensed service is black-letter regulation. In 2004, the FCC asserted that it has the sole authority to resolve interference in the unlicensed band. The FCC is studying the scuffle at Logan Airport, and I hope that they uphold their previous rulings. Massport has been using spurious arguments about interference and the need to control the network for optimum performance due to the critical applications they are using. That may be true, but in that case, they should be using the 4.9 GHz band, which is licensed solely for public safety usage, and therefore more appropriate for key card access and police communications.
-
802.11 is unquestionably the last-hop technology for most users. It’s spread well beyond laptops and PDAs to include VoIP telephones and even consumer devices such as streaming media servers and digital cameras. It’s too bad that most of the consumer devices don’t yet implement real security protocols and stick you with antiquated security based on manual WEP keys. As 802.11 chips get cheaper, it will become a more common transport medium for home devices.
-
The task group working towards an 802.11n standard has yet to take the first key step. As the year opened, there were two draft proposals, WWiSE and TGnSync. We end the year without having selected one, but at least there were some interesting meetings along the way, and the emergence of the proposal from the Enhanced Wireless Consortium. A year ago, I was putting the finishing touches on the second edition of 802.11:TDG, and wondering how long the information would be current. Given that we still don’t have a draft proposal, I guess I can feel pretty good.
-
After task group approval in July, 802.11e was finally ratified in September. It started out life as both QoS and security, but the security components were split off into 802.11i. (That security was considered “easier” should tell you how difficult QoS was.) I have always been a bit of a quality-of-service skeptic, in part because QoS through over-provisioning has almost always worked on the LAN. Wireless LANs have much more limited bandwidth, and the use of the less power-hungry 802.11b further restricts capacity. In the case of 802.11, I expect that 802.11e will be a good start, but there is room for further QoS enhancements for extremely dense deployments.
-
The final item is not so much a news story as the lack of a story. 2005 was a quiet year on the security front. There were no major new stories about flaws in the protocols, and standards development has slowed down considerably since the ratification of 802.11i in 2004; what remains now is to make all the implementations play nicely together, and get them widely deployed.
