Dan Keller's Wifi Links 'n Stuff

Latest news (2008): We have given up on SFlan (see below) -- never did get the three-radio setup completely working -- and gone with Meraki ("Free the net") who is doing what Google and Earthlink couldn't.
The economics of wide-area wifi -- like the technology -- has yet to shake out. Neither one is yet stable or mature.

Perhaps a newish technology, 802.16 (metropolitan area networks, also called WiMax) will win in the end... but don't bet on it. Don't bet on anything just yet!

According to this article, "True wireless metro area networks remain the holy grail of the wireless industry." The solution will be a combination of cellular, wifi, and wimax -- a kind of cellular LAN, which is what the sflan people have been doing.

What business opportunities does wifi offer? Who knows? It's changing so fast! All you can do is hold your nose and jump. The shakeout is really happening; spectacular failures like Cometa and Toshiba reveal that wifi revenue will not come from the end user. Rather, it is a retail amenity, given away to sell more coffee, ice cream, hotel rooms, apartment rentals, and trailer park spaces. The revenue comes from housing, restaurant and hospitality vertical markets. The key is to add value, provide real service, and be scalable.

So that's my strategy. Stay tuned.

SFLAN is "...an experimental wireless community network in San Francisco. We aim to build a wireless network with LAN characteristics on a metropolitan scale." Extensive coverage has already been achieved with fifty nodes in place and more on the way... including sflan50 on my rooftop in San Francisco's Castro district. We're lighting up Market Street! Yee-hah!

Inside that weather-proof box is a Linux machine with a flash memory "disk" and three radios, all powered by PoE (Power-over-Ethernet). We manage it with wdistconfig. Uptime is measured in months.

Here are more pictures. The other guy is techno-visionary Ralf Muehlen. Here he is at the April, 2006 Maker Faire.

San Francisco has had various failed attempts to build a public, city-wide, "free" wifi network. An early effort, spearheaded by Mayor Gavin Newsom, died when the accepted bidders (Google and Earthlink) dropped out. Read about it at Wikipedia: San Francisco Municipal Wireless.

SFLan, too has been a disappointment. It had a modicum of early success but, as Ralf has written:

SFLan, a project of the Internet Archive, constructed an experimental cooperative wireless internet service provider (WISP) that grew to approximately 30 nodes, using an over-the-air backbone relying on line-of-site transmission. Noise level in the ISM bands due to proliferating and competing Wi-Fi signals made many of their long distance links (several miles) nonfuncitonal. The network contracted from 30 functioning nodes in February 2007 down to four by October. They are considering changing their architecture to a fiber and wireless hybrid.

Nonetheless, we like SFLan's approach best:

The point of SFLan is to demonstrate that it is possible to build a third pipe independently of the phone and cable duopoly, with free software and commodity hardware, and without any strings attached. That means a neutral network, with no ads, no censorship, and no restrictions on privacy.

In general, the big issues are:

Here are some resources for people working with wifi:

* CNIT: Computer Networking Information Technology at the City College of San Francisco