My cell phone is a Nokia 3360. It came gray, but I spiffed it up with a yellow case that probably cost about as much as the phone -- to buy, that is; not to make. Your big cell phone margins are in aftermarket knicknacks like earphones, replaceable plastics and downloadable ringtones.

Nokia sold upwards of 81 million phones in the 3100 series. Plus another XX million

Here are a few numbers from Gartner:

Newsweek says there are 1.5 billion cell phones in use right now.

How do you get mobility and decentralization?

When I connect with a smartphone over gprs, my phone gets an IP address, just like any other Internet device. I can run a small http server. I can do cahat IM, and even ssh.

Back when Transmeta announced the Crusoe chip, they showed off a bunch of prototypes... one was an erector set.

here's the problem:


Russell Beattie: Mob Law

Works for Wavemarket, has a blog.

Mobility is growing like crazy.

Wi-Fi is really cool and interesting, but it will never touch as many billions of peopl e as the cell phone already has, and will continue to touch over the next three or four years.

"Docomo and KDDI are selling wireless internet access to customers for a flat rate. That will spread to Europe and the U.S. as well."

"Hutchinsons 3 tried the walled garden approach, with 3G services that's in a walled garden. Abject failure. An object lesson.

Now they're competing over services, not controlled intellectual property. Even microsoft will have to compete on support for services.

When I connect with a Nokia Series 60 phone over AT&T wireless, I get an IP address, and it's not blocked at all.

For Linux the question is which of the smart phones will win. There's Symbian, Linux, Microsoft and

Just for cost reasons, guys like Sanyo, Samsung, want, just for cost reasons, not to have to create custom hardware controllers everytime theycreate a new phone. There's a need for generic hardware and software. Walled gardens are discredited.

On the handsets there will be lots of room for Linux. Phones coming out of China...

Me: the value system of Linux is starting to prevail.

With Sumbian on a Nokia 60 phone - pentium processor, 32 megs ram, a real 32 bit OS, a file system, and an internet connection, I was able to take pix, ftp them to my web server,

A platform in my pocket

Symbian released its OPL language source code under the LGPL last year.

Go to Symbian page. http://www.symbian.com/technology/symbos-v8x-det.html

This is what Linux is going to have to come up with. Either the Linux community or the cell phone manufacturers, or both. The questions are, "Is it worth it ot me to come up with my own CDMA implementation? Or is it worth the license fee to use Symbian?" I know there are a growing bunch of libraries out there. Bluetooth, for example.

Motorola came out with the first Linux smart phone last year: the A760. [Note: high-res images here and here. But probably need to get Moto permission first.] No mention of linux in the copy. It's just a phone.

"They're just throwing spaghetti against the wall right now. They have Microsoft, Linux and Symbian phones, in additoin to ones with their own OS."

It suports Java, a camera, bluetooth, an MP3 player, PIM functions, WAP and a browser.

Linux Devices on the same subject. Many links there.

Also the E680. Rundown here.

More here about other Linux cell phones.


Even low powered phones are going to have Java, and memory, and net access. IN countries with relatiely little broadand. These people are going to get on the Net, and it will be a huge change.

What will change the world won't be a Simputer or a cheap Linux box. It will be volume pricing on cheap Net-ready cell phones — maybe running Linux.

I want a GPS phone. Where am I?

Wi-Fi requires a lot of infrastructure. The cell phone system will mostly already be there.

Wi-Fi is cheap parking for Lexus drivers. Mobility is free parking for everybody.

There is a custom in Spain on holidays to send out SMS messages to everybody in your address book. Something like 30 billion messages were sent this last New Years.

in the U.S. there are 40 million DVD players and 100 million cell phones.

what matters is communications itself.

Carrier Grade Linux: Note who belongs, recent news.

Tim Pozar,

Now technical director at the EFF in San Francisco. Transitioning from the tech department doing just IT, to the techical department actually doing something disruptive. A bunch of projects under the technical and legal departments. Some tech, some legal, some both.

Developing a twirl-like anonymizer network called spartacus. Another is demonstrating to the FCC that we don't need a broadcast flag for DAB. Another is a white paper on how to make ISPs ad WISPs clia? and DMCA safe. In other words, how to make your logs disappear quickly, so you don't have anything to hand over. We're doing a patent busting projec t that was announced. We're doing HDPDRs, so people an build their own before the broadcast flag gets implemented into ?five. We're going to be working on spectrum inventory, so we can figure out what spectrum is actually going to be used in various real-world situations.

There are two protocols. one for fixed, one for mobile. (Wimax) Dwayne's list. article that talks about the dfferences between 802.11d and e. Break it down to one is more designed for mobility, the other more for fixed. I'm hoping that the fixed does work.

wifi is the compliance org. 80=2.11 is the protocol name.

For fixed, long distance stuff, 802.11 doesn't scale. It works, kind of.I'm hoping that 802.16 can get to the commodity prices of 802.11. it probably won't because it won't be as popular and therefore won't commoditize. You won't see it deployed in offices, for example, while 802.11 works there.

The hard part for us is getting mountain tops, so we can deploy whatever the eventual solution might be. Have an access point at UC berkekely. Approvals for KPFA (Grizzly Peak) and KFJC (Black Mountain). We is BARWN, the Bay Area Research Wireless Network. Deploying with Soekris boxes.

We want to make sure this network does not go the direction that the rest of the INternet did, which is asymeterical. The biggest objective is to deploy this stuff so economically depressed areas can deploy the same technology. It's easier here because we have Fry's down the street.

Designed for fixed links. We distribute through neighborhood access points. We're trying to address the second-to-the-last mile issue, due to the fact that broadband is usually encumbered, so yu can't share it. We supply the broadband. Do whatever you want with it. If you went to the phone company and said "I want to share out my DSL connection and see what happens.

Check Jon Katz, working with Cornell to deploy stuff in Nicaragua.

Also Bhutan, where, withthe help of Clif Cox, they're "leapfrogging the need to spend lots of money for carrier class equipment by putting in 802.11 and doing VoIP over it."



On IBM: Building a wireless access point on Linux. Linux On Soekris has a lot of info, too.

BAWUG, the Bay Area Wireless Users Group

Developing a

Matt Peterson:

Founded BAWUG, and tim's "co-partner in crime on the BARWN stuff."

Place radios around the city and monitor spectrum usage. More in house projects, more proof of concept. I'm reallyinterestedin the VoIP stuff.

Gave a presentation in Kuala Lampur. Asia has no NANOG. Instead they have Apricot. They hae APNic, which is not commercially driven. They're a real nonprofit. Apricot is a yearly conference. They focus on education and best practices — how do you deploy wireless, etc.


the media tends to wrap these up. In reality they're very different. WISPs are mostly in the midwest, and internationally. Domestically, they're a lot like the entrepreneurs we had in the ISPs we had in the mid-90s. They're willing to try new technologies, to crawl around on roofs, to buy Tupperware and put wireless equipment in it.. They're into their user base.

Hotspot operators are mostly aboutenabling business travellers to get reliable internet at Starbucks. Lots of fluff going on there. If you don't make chips, you don't make money in this industry, and that includes products, software, integration... It's very difficult to make money.

T-Mobile is basically a phone company. They bought out MobileStar. They deploy carrier class gear. Minimum $3000 per location.

Cisco switch, AP, router, and a Cisco hotspot soution that's also fairly expensive and very configurable.

Wayport has a compbination of typical network equipment plus a Linux box and a custom portal. Also T-1. In the beginning they did conference centers and hotels, with wired ports. They deployed 802.11. They gave it away for free for a number of years.

Surf & Sip brings in DL, because there actually isn't high demand for bandwidth. They also use an embedded machuine running BSD. hardware costs are $200. They drop-ship it. Much different than ligting up a starbucks, where you'e spenindg at least $3000 and have to work with the telco.

There are approaches like NYC Wireless, where you rebrand your access point as NYC Wireless. "But that just extends the monopoly."

Now Surf & Sip is turning into what Cometa wanted to be, which is a an outsource for build-outs. C

It's hard for the mom & pop shops.

I've always been in the middle. When you have a Mom & Pop, and plug in a Linksys AP, usually it's the worst possible situation. One, i'ts illegal, tpically. The AUP doesn't allow it. That's beena problem in New ork the cable company drives around in vans, looking for violators. This has changed int he last year. Now the cable companies give you a router hat ha siwireles but in , and they charge ou for that. That's what they've always wanted: to control as much of the edge as possible. Second, you've got an access point that technical folks know how to use. What does the barista say when the customer says they can't send email, because port 25 is blocked? You can't say "wiat until the geek comes here next week."

There's a model like Emenitiy's, where you give or lease a cheap managed box you can SSH to. Charge a $50/month management fee. Not easy to find a business model here. But Ultimately this will be most successful.

Copmmunity wireless networks dependon an endless supply of geeks. They get burned out. They don't want to do it after 2 or 3 years. They were looking for a utopia of plaing ames and tading MP3s. Not of extending broadband. It's a tough choice for these organizations. Should you extend Speakeasy or build your own network? "I'm against extending monopolies. I did it for a living. There's nothing exciting about it, and it doesn't change anything." It has a lot of baggage. If you build a nework yourself, you can gget rid of all that baggage, those legacy rules. Hey, I want to play with RIP Routing. Okay, it doesn't work. It floods the network. Want to play with IPV6, go ahead. The great thing is , you build a great wireless netwolrk and start connecting homes, and go downtown and upthe hill, youeventually can say to the Mom & Pop cafe, 'buy this linux box, with one wireless card up on the roof and one in the box, and you're set. That's exciting. I only know of two networks aggessively doing that. One is in the Netherlands, in Leiden.

Surf & sip did it cheapest.

Reese Jones: The 100-year advance

software for telephones.

Jim Thompson: Consider the Source

Nigel Ballard


James Ewing

David Isenberg

Kevin Werbach

David Sifry

Glenn Fleischman

Britt Blaser

Deus ex mobilia

Doc Searls is Senior Editor of Linux Journal. He also writes SuitWatch, a bi-weekly Linux Journal newsletter. He also presides over Doc Searls' IT Garage, a sister site to Linux Journal on the Web.