My initial introduction to wireless networking was in Monterey, California, at OSCON (Open Source Conference) 2000. O'Reilly arranged free public wireless access for conference attendees. The tremendous flexibility of being able to connect to the network from anywhere led to all sorts of interesting, unforeseen interactions. For example, people attending a large talk could converse in real time over IRC and discuss the talk (and even critique the speaker) without raising their voices. They could use the Net as a resource when asking the speaker questions, to draw out very interesting points by way of real-time examples. With an instant messaging client, ubiquitous wireless made an effective, free, two-way paging system. (Rather than trying to use the overloaded PCS phone system, it was now possible to send a quick "Where do you want to meet for lunch?" message and get a response back instantly.) Conference attendees no longer had to return to their hotel rooms for dialup access, or be banished to a terminal room away from where the action was, just to check their email or refer to a web page. That was assuming, of course, that one had an 802.11b card and laptop handy. Personally, I had to wrestle a card away from a buddy who happened to have a spare. I realized that networking on borrowed time wouldn't cut it; I simply had to pack my own.
On returning from OSCON, there was much interest at O'Reilly in getting wireless networking going at the office. If that much flexibility could be put in place for very low cost, why weren't we using it in-house? If conference-goers could use the stuff to grill speakers for information more effectively, what could it do for our company meetings and presentations? And so, without even knowing my Direct Sequence from my Spread Spectrum, I started down the long, winding path of wireless networking.
9.1.1 The CampusAfter setting up a couple of access points to cover our campus, and a crash course in WEP, MAC filtering, and closed networks, our fledgling 802.11b network was up. With relatively little effort and expense (about $3000 and a few hours work in all), we now had seamless coverage in all three of our buildings, complete with roaming between APs. The main O'Reilly offices in Sebastopol consist of three two-story buildings, covering an area about 450 by 150 feet. Using one Lucent AP-1000 in each building, and a small 5db omni at each AP, I was able to cover nearly all of the offices and conference rooms.
Early on in the process, one of our users noticed that they couldn't get online, even though they had a very strong signal. Upon checking their network settings, I realized that they hadn't set their ESSID, and so were associating with any available network. It just so happened that the network with the strongest signal was coming from the business next door! I fired up Lucent's Site Map tool, and, sure enough, there was an existing 802.11b network immediately next door. After a quick conversation with their sysadmin, we decided on a channel numbering scheme that would minimize interference between the two networks. (This is exactly why a preliminary site survey is so important: even though you may not see antennas, a network may already exist in your area! Don't just assume that since wireless is new to you, it's new to your part of town.)
Now that our offices were saturated with access, with 50+ users up and happily untethered, what could we do with it next? Naturally, more than a few eyes turned to the hotel and coffee shop across the street. If one could get a signal from the hotel, then visiting employees who stayed there could get online for free, at 11Mbps (as opposed to paying for a trickle of dialup access). And of course, being able to work directly from the coffee shop must do something for productivity. With visions of mochas and bandwidth dancing in my head, I looked into adding external antennas to increase our range.
In about a week, I had an omnidirectional antenna installed on the roof, running down 25 feet of LMR-200 to our access point. Why did I use 25 feet of cable that loses almost 17db of signal every hundred feet? And why did I use an omni, when a tight sector or patch antenna would have made more sense? Because a year ago, without any prior background in radio, I went with what our vendor had to offer: a 25-foot run of so-called "low-loss" microwave antenna cable and an expensive omnidirectional popsicle stick. (After all, if Lucent made the gear, it must be compatible with a Lucent access point, right?)
Luckily for me, even with the high line loss, the omni managed to do the job. That afternoon, I walked across the street, ordered an iced mocha, and merrily typed out the confirmation email. As I hit Control-X Y, I was compelled to meditate on that inevitable question, "What next?" If it was possible to get a good signal about 1500 feet from the AP, how feasible would it be to provide wireless access to our local employees? After all, many of our people live in the area and were using dialup to access our network from home. Would it be possible to provide a fast wireless connection to anyone who was within range? Just how far could this technology be stretched?
9.1.3 Online from Home, No Strings AttachedAround this time, I relocated to Sebastopol from San Francisco. By a staggering coincidence, the house we moved into happened to have a clean line of sight to the antenna I installed on the roof, more than half a mile away. This provided a great tool for experimentation, as I now had a fixed signal at a distance with clear LOS and could aim whatever kind of equipment I liked at it to see how well it would perform. I realized that a high-gain dish, pointed directly at the omni, could achieve a very good signal, even through walls and glass. I was so excited by the quality of the signal that I bungee-corded the dish to a chair with rollers, and rolled it around the house, while streaming a full-screen video on my laptop the entire time. Yes, a keen interest in wireless was now developing into a full-blown psychotic obsession, as the potential possibilities of long-distance, low-cost, high-speed communications played about in my mind.
I finally mounted the dish inside my attic, set up a makeshift access point, and found that I could have a stable 11Mbps connection from about six-tenths of a mile away, with a "stealth" dish under my roof that wouldn't bother the neighbors. I used this connection for several months, through all kinds of weather (and I was very grateful of it: in my area, DSL and cable modems weren't an option at the time).
Now that I had a proof-of-concept and parts list, I approached others in the company who live in the area, to try to set up a second node. This was when I came up against possibly the biggest natural obstacle to long distance microwave: trees. As it turned out, I had been truly lucky with my own situation. Finding many clean paths to a single point is highly improbable in Sebastopol. Except for the immediate downtown area, medium to dense foliage is virtually everywhere. After visiting several possible node sites (and trying to shoot to O'Reilly despite the trees), it became clear that a single access point at a low altitude wasn't going to be sufficient to get our Sebastopol employees unplugged. There are just too many trees between O'Reilly and the rest of the world.
With no obvious plan of action, I had to put the wireless extension project away for a while so I could do more research. By now, there was certainly no shortage of online information available, as community groups began popping up all over the globe. I decided that if I was going to get anywhere with practical wireless networking, I'd need to talk to some experts.
9.1.4 Seattle WirelessLast March, I took a trip to Seattle. My brother was moving to the area, and so I took the opportunity to travel with him up north to see the Seattle Wireless network for myself. I must admit that wasn't fully prepared (psychologically) for what I found when I got there. Here were a bunch of very sharp sysadmins, programmers, and net monkeys, who were gearing up to build a redundant, fully routed public network, independent of the Internet. They were working on this project entirely in their spare time, with no promise of reward other than the joy of hacking out a project that simply needed to be done. They weren't just hooking up a couple of APs and trusting their luck; they had an entire network topology planned, a hardware solution down, and nodes in the works to connect sites miles apart.
I spent a day building antennas and speculating about the possibilities of 802.11b with the SWN crowd. By the end of the day, we managed to put together a yagi made out of washers, some tubing, a bolt, and a pie tin that carried an 11Mbps signal about a mile. The topology of Seattle is such that their network plans will probably work: tall buildings, rolling hills, and limited tree cover make much of the city accessible (assuming one can get on top of the hills). I went back to Sebastopol with a few important realizations:
• There was tremendous interest in high-speed wireless networking, even among people who already had high-speed wired access. Ubiquitous wireless seemed to be almost as much in demand as DSL and cable modems.
• The seemingly insurmountable difficulty of finding LOS between points can't really be approached by one person or group. But a larger community, working together toward the same goal, can bring a lot of resources to bear on any problem.
• Wireless networking isn't as simple as replacing a piece of CAT5 with a radio. Radio has many strange properties that are completely alien to people who have been studying computers and networking for years.
• Conversely, many radio experts find themselves lost when dealing with the intricacies of Internet networking (until very recently, a 9600-baud packet radio connection to a computer running a DOS TCP/IP stack was considered high tech in many circles). If we intend to push 802.11b beyond its intended limits, the plateau of knowledge that separates hardcore network jockeys from hardcore radio geeks must be crossed.
Ironically, it started to look like it would be easier to get the entire Sebastopol area unplugged with open network access, rather than trying to connect a few users to a private network. But to do that, I certainly couldn't do all of the work. I needed to find out if there was as much interest in my area as there seemed to be in the rest of the country.
9.1.5 NoCatIt was obvious that we needed a local repository for information about forming a cooperative community network. Within a couple of days, some friends and I put together a simple web site and mailing list. But what to call it?
While sitting on the couch in the living room, logging in to check on something at work, my login fortune struck me as particularly funny, so I read it aloud to my friend Cat:You see, wire telegraph is a kind of a very, very long cat. You pull his tail in New York and his head is meowing in Los Angeles. Do you understand this? And radio operates exactly the same way: you send signals here, they receive them there. The only difference is that there is no cat. -Albert Einstein
Cat quickly replied, "That's what you should call this thing: No Cat." I immediately checked whois and saw that nocat.net was free. That settled it.NoCat became the central repository for several wireless projects that Schuyler Erle (Perl programmer extraordinaire and wireless sympathizer from O'Reilly), myself, and others had been working on. We put together WRP, a wireless router-on-a-floppy to make setting up a wireless gateway quick and painless. We also started work on the NoCatAuth project, a method for authenticating users to a cooperative network without using any of the built-in (and limited) authentication methods available in the 802.11b spec. We also set up a mailing list for locals interested in wireless. Now that we had a web presence and some information available, we needed a way to connect with people in the local community.
9.1.6 The ArticleLuckily, we didn't have to wait very long for the community to notice us. Just after I returned from Seattle, a local newspaper, the Press Democrat, (see Archives of http://www.pressdemo.com/ for "Downtown Sebastopol Going Wireless," published on April 2, 2001.) ran a feature on some of the wireless shenanigans I had been up to. I had no idea at the time how valuable this kind of exposure could be to the community LAN idea. Within a week, I had received a few dozen emails and several phone calls from locals who were interested in wireless networking. Some offered expertise and equipment, while others were simply curious about our plans and what could be accomplished with 802.11b.
After the article ran, our mailing list grew to about 25 people. We decided to hold a general meeting to get organized and figure out what we wanted to do with this stuff. I was very pleasantly surprised when 16 interested people showed up at that first meeting. Many were looking for free high-speed access, while others were simply curious. A few were Northpoint victims who had been forcibly unplugged from their DSL when that company went under and who were looking for anyalternative (apparently, they were no longer considered part of the "prime" market and would likely not see high-speed access again for quite a while).
As the discussion went well into the third hour, it was obvious from that first meeting that this was going to turn into a regular event. These people were keenly interested in contributing to a free local network, and they had a tremendous amount of knowledge and resources among them. But until now, they had no good way of connecting with each other. From this first gettogether, all sorts of possibilities began to present themselves.
The general consensus was that, if people who had high-speed Net access wanted to share with those who wanted it but, for whatever reason, could not get it, there were several technical obstacles that needed to be overcome:
• The solution couldn't cost thousands of dollars, or else no one could afford it.
• There had to be a secure and easy way of figuring out who was who and limiting what users could do on the network (so that node owners wouldn't be exposed to abusers or have their hard-earned bandwidth monopolized by a freeloading few).
• The solution needed to be simple enough for someone with limited skills to set up, and it needed to require little or no maintenance.
• There had to be an easy way for people interested in point-to-point links to meet with each other.
• People who did have a fair idea of how to proceed needed access to all sorts of information, from choosing microwave connectors to configuring laptops.
We had some answers to these issues, but it became clear that these were going to be longterm problems, shared by anyone attempting to put together a community group. We put as much information as we could up at NoCat and pointed to others who had answers whenever possible.
The nearest wireless community group to Sebastopol was the BAWUG, who met regularly in the San Jose area. Since we were obviously working along parallel lines, it seemed to be time to see what our neighbors to the south were up to. I got a couple of local wireless zealots together, and together we made the two-hour trek to San Jose for a meeting.
The June 2001 BAWUG meeting was a great opportunity to network further. Much like our Sebastopol meeting, there were people with all different abilities and expectations present (only here there were about fifty of them!). After a couple of interesting presentations, I got a chance to talk with antenna gurus, some Apple AirPort hackers, and even a commercial wireless startup.
There was much buzz about the impending Portland Summit meeting. Wireless community leaders from all over were going to converge on Portland for a weekend of planning, talk, beer, and general hackery. This was a meeting I could not miss.
9.1.7 The Portland SummitLast June, for the first time ever, people from community wireless networks across the country (and even from Canada!) met in Portland to talk about what we were up to. Organizers from Seattle, New York, British Columbia, Portland, the San Francisco Bay area, and Sebastopol were there. We had a very productive couple of days, covering divergent topics such as antenna design, network layout, the FCC, and "catch and release" captive portals. There was a tremendous energy and goodwill between the groups, as we all realized we were in this experiment together (admittedly, the beer probably helped a bit, too).
I think the Portland Summit was very reassuring for all of us. Here we had people from all over the globe who share something of a common vision: unlimited free bandwidth everywhere. We developed these ideas independently, and while some of the details of how we were attempting this feat diverged, the ultimate intent was the same. Wireless community network access seems to be an idea whose time has finally come; hardware, software, and network backbone are all becoming cheap and ubiquitous enough to make it happen. All of our groups want to strengthen our local communities by bringing network access to anyone who cares to be a part of it. And by working together, sharing what we've learned, giving away software, and pooling our collective efforts, we are finding that we can reach this goal faster than by trying to work out a solution on our own.
9.1.8 The FutureNow, with about 150 people on the mailing list, regular monthly meetings, and an ever increasing number of public access nodes, we are definitely moving toward ubiquitous, free network access for whoever wants to use it. The NoCatAuth software has been officially released (check out http://nocat.net/ for details), and the priority in Sebastopol is now to get nodes online.
With cooperative effort and wireless technology, the Internet is rapidly becoming more and more pervasive. My direct experience with people working on this project has turned up an important common thread: free access to information is in constant demand, and barriers to that access cause pain. I believe that working to provide free, unrestricted access to the Internet is a benefit not only to one's local community, but also to the world at large. I hope that this book has helped you to realize your goals and has helped you become more connected to your local community.