May 5, 2006 4:00 AM PDT

Free Wi-Fi in S.F. more than flipping a switch

Tony Bennett left his heart in San Francisco. Will EarthLink and Google leave a barrel of money trying to deliver universal Wi-Fi access to the hilly city's 740,000 residents?

San Francisco politicians, like many of their counterparts in cities throughout the country, view affordable broadband access for all citizens as essential, like providing water or electricity. In October 2004, San Francisco Mayor Gavin Newsom announced that he intended to provide free wireless access for the city.

He's trying to fill a gap in the city's existing, piecemeal Wi-Fi network. San Francisco is already filled with hundreds of smaller Wi-Fi hotspots offered through coffee shops, retailers or neighborhood groups. But Chris Vein, executive director of the city's department of telecom and information services, said that the goal of the citywide Wi-Fi project is to bring free "universal" Wi-Fi to residents.

"The Wi-Fi hot spots today are not covering all 49 square miles and don't reach out to poorer areas," he said in an interview with CNET News.com.

In April, the city announced that it had selected Google and EarthLink to build, manage and provide service over the network. The companies have laid out a plan that is expected to meet the city's demands for ubiquitous and low-cost access for nearly 90 percent of the population. And they say they can do it for $15 million--including maintenance and upgrades--in the next 10 years.

Critics wonder: Is that possible in a city with downtown high-rises and a 900-foot hill smack in the middle?

The EarthLink/Google plan includes two main options for broadband service: One allows subscribers to pay about $20 per month for a 1Mbps connection from EarthLink or another Internet service provider leasing capacity on the Wi-Fi network. The second option allows anyone to get 300Kbps download service for free in exchange for viewing local advertisements through Google.

Click here to Play

Video: Ups and downs of San Francisco
Google and EarthLink have partnered in a project to bring universal Wi-Fi to San Francisco. Here, a view of the hilly terrain challenging the install.

EarthLink said it expects the project to run to between $6 million and $8 million in initial costs, which include attaching radios and receivers to utility poles throughout the city. Within 10 years it expects the whole network, complete with upgrades and maintenance, to cost about $15 million.

Finer financial details of the project haven't been made public, but the plan calls for EarthLink and Google to contribute to the initial cost of building the network. It's not clear what the split between the two companies will be. Once the network is built, Google will pay EarthLink for access to the network on a wholesale basis. In order to make access free to people in San Francisco, Google will use revenue generated from local advertisements to pay for access to the EarthLink network.

San Francisco may well be one of several test markets for Google and local ads. Finding local information over the Internet is a highly fragmented business, divided among search engines, interactive yellow pages, local business sites, wireless carriers and so on. Google wants to pull those strands together and make targeted, useful ads a reality.

Ad test market

Google and Yahoo are after the same local ad dollars. But the two companies are making different bets as to how consumers will access digital data ultimately. Yahoo, for example, has teamed with local phone companies to be the default search destination when someone signs up for the $19.95 Digital Subsciber Line (DSL) service. Meanwhile, Google is teaming with EarthLink (and likely others) to be the default provider of ad-supported Wi-Fi. Its projects in Mountain View, Calif., and San Francisco are likely be just the beginning.

While the Wi-Fi plan sounds compelling to many, some critics say EarthLink has underestimated the cost and overestimated its ability to reach most of San Francisco's residents. Others are skeptical that Google will be able to sustain its free model with advertisements alone.

Among the technical concerns are San Francisco's famous hills and tall buildings that get in the way of radio waves. Just finding adequate cell phone service can be a hassle in certain parts of the city.

Don Berryman, executive vice president of EarthLink in charge of municipal wireless deployments, acknowledges that deploying Wi-Fi in San Francisco will be tricky.

"I would never say that San Francisco will be an easy place to deploy Wi-Fi," Berryman said. "There will be a number of challenges. But I think most of the critics aren't looking at today's technology. We aren't talking about using Linksys routers to set up a hotspot. These are high-powered radios that send signals 600 to 1,000 feet."

CONTINUED: Wi-Fi takes to the hills…
Page 1 | 2

See more CNET content tagged:
EarthLink Inc., San Francisco, city, Wi-Fi, Google Inc.

7 comments

Join the conversation!
Add your comment
Build out will kill EarthLink
Earthlink is already underestimating the number of Mesh Nodes it will need to cover the City, especially the downtown area where usage will skyrocket. Tempe Az required 650 alone and it is smaller than SF and flat too.
The other problem they have will be the number of Gateway links (Motorola Canopy radios)they will need to keep the bandwidth (in the single radio Mesh Node) at a reasonable level to retain users. All I can say here is Motorola will love the business it will get from deploying a Canopy Gateway at every 3 Nodes. Same problem Earthlink will have in Philly.

Will be very interesting to watch this develop.

The industry analyst need to start commenting on this so the entire Metro Area Wireless market doesn't get a bad name because someone made a mistake in picking the best hardware.

Jacomo
Posted by jacomo (115 comments )
Reply Link Flag
Agreed - this will cost 3X estimates - easily
Metricom built out San Francisco with reasonably speedy and
extremely useful wireless networks twice. Once in the 90s, and
again in 2000 with a high-speed 128kbps network. (Does no
one at News.com do any research when writing these articles, or
is comparing current efforts with previous failures not germane
somehow?)

What did Metricom learn? Even when using two or three
frequencies, radio densities were much higher than forecast -
and so was the cost. Reflection, multipath, and other
interference was higher than expected. And this happened to a
company that had already done the task once!

Metricom was able to build out seven metro areas with high
speed wireless for a little less than a billion dollars. Earthlink is
crazy if they think they can do the same thing with Canopy and
other "bolt on WAN" technologies for 6-8 million total.
Metricom's mesh network architecture was built from the start
for redundancy and survivability, enhancing frequency reuse,
while 802.11-based products can't fully escape their original
purpose - channel-based static access points. They're far more
inefficient than Metricom's old architecture, even with advances
in computing power and RF processing.

First you need to survey everything. Just getting municipal
clearance from all the agencies involved with lightpoles in S.F.
will be a nightmare in and of itself. Then you have to hire bucket
trucks to visit most blocks in San Francisco to hang radios. Then
you need to survey everything again. And then you can start
working out all the bugs, weak spots, dropout areas, and
addressing the vandalism and other maintenance issues. (Lots of
birds in S.F., and two things attenuate R.F. like nobody's
business. Bird bodies and bird...*hit.)

This all seems to point to a number far higher than the 6-8
million Earthlink is hoping for. I have to wonder who there is
doing the estimating.

Earthlink was interested in Metricom's technology, too, but bad
marketing and management doomed the company to failure
during the dot-com crash, and apparently Earthlink caught the
death smell on the company before investing.

It'll be interesting to see how successful this venture is - but I
think it'll end up costing all parties far more than forecast, it'll
be disappointing when finished, difficult to upgrade in the
future, and it won't help bring any poor people the magic of the
web, either - unless Gavin and his new pals find someone to
subsidize that little project, too.
Posted by Hep Cat (440 comments )
Link Flag
The Idea is there, the means however.....
I think its a great plan, I would love the accessibility in my city, as a matter of fact I am going to plan it out and present it at our next city counsel meeting. However in a city like SF there is definately a lot to consider. The cost of course is an issue, however if it does get to be a little much they can still decide to charge a small fee, 170,000 people paying $10 a month minimum can definately keep it running. Besides with a business city such a SF there is a definate need for all over access. There are still more issues that need to be focused on before this will even work. For instance, interference. My phone alone knocks out my wireless. Hackers, my god how much would they love access to all of SF. Either way, I think with more planning, acceptable hardware positioning and enough backing to pay for it, this idea can actually take off and expand nation wide. Besides, with rising gas prices I can barely afford to go war driving anymore.
Posted by Mechy20 (4 comments )
Reply Link Flag
Breakfast at "Tiffany's"; coming home on the 'Red eye'.
Jet lag is only a raw egg and a shot of 'Jack Daniels' and I'm fresh for some talkback with 'The (DBA) Critics' Redmon review. #1 topic of interest from my hometown Jewish Deli; 81% of our businesses are run from the 'Home' in this community. Odd as it may sound, in a Republican state like mine their's a whole lot of money in the private sector. Translation: Everyone is vested in Keo's and 401k's. More over, I've paid for my personnal health insurance since 1998; even as a contributory expense. But, hey I was employed doing what I had done my entire life thow this time in the organic produce department at "Wild Oats ranch market". My history is shallow to most going from aquisition to stock split and back again. But, this municiple wi-fi has potential; with the IPO's at heart, like Cica-Cola, Layes Potatoe Chips and (maybe a cash account business such as Budweiser). They all need broadband at this very time of perqyerment.
Posted by Pop4 (88 comments )
Reply Link Flag
The Story Says they Plan to Spend $15 Million ...
before all is said and done, but I think they meant that would be what was needed for a first phase of deployment and operation - the $6 - 8 million is probably just an alpha/beta initial capability in the areas of highest payoff - where people are mostly likely to want to use this capability beyond the existing capability provided by Starpukes, McDumbalds, etc. (and that would help keep the wee-wee contests with the corporate types to a dull(er) roar). I do agree that it will take far more than even the $15 million to be able to provide free access to every homeless guy's cardboard box around the city(just think about what kind of propagation/reflection headaches all those shopping carts full of aluminum cans are going to cause! :) )

However, remember that there is a paying option, and you can bet your sweet bippie that those customers will have a shiny (although, eventually, bird-poop covered ;) ) new access point within eyeshot of wherever they use their laptops. WiFi technology is much cheaper to field, even in the hundreds (thousands?) of nodes than cellular technology, especially if it's mesh-capable. Unlike cell phones, most current WiFi devices will not be used across nearly as much real estate (especially in mobile mode where the user is in control, e.g., walking, driving, etc.) - and anyone who does attempt that will soon be cleansed from the shallow end of the gene pool (but, why is it always _after_ they've reproduced?) by a cable car, trolley, bus, taxi, cyclist, mugger, etc. WiFi phones won't be a problem, as long as they are multi-mode and can take advantage of the existing cellular infrastructure (anyone know whether they default to WiFi when it's available? That will put additional strain on the network, if/when they start showing up in large numbers, but I hope for Earthlink's sake that it doesn't happen for a few years down the road/hill).

Other than proximity to Google's HQ, there is one very smart reason to start with San Francisco as an early nut to crack - if they can make this work there, it will be a relative breeze to make it work anywhere, even the glass and steel canyons of New York City. The continuing decline in the cost of the technology (and massive buying power) will also mean that cities they attempt this in elsewhere will be increasingly cheaper to do, from both hardware and labor costs (I'm guessing that propagation/usage models will be refined over time that will make each nth installment easier - but, still not necessarily easy).

It is curious that Cisco/Linksys has been silent about all of this commotion - maybe they're planning to buy/team with Yahoo/AOL/MSN to do the same thing in other places (or, maybe they're waiting for Google/Earthlink to prove/disprove the concept, and then do a me-too, or skip the mistake, in cities not already slated for this kind of capability. I wouldn't count out Microsloth leading its own team, either, with the obvious first deployment being Seattle (which may be third only to San Francisco and the sprawl of LA in geographic difficulty. Cities like Omaha are not going to be nearly as much of a challenge, now or in the future (and there won't be nearly as much money to make there, either, so don't expect a big rush of companies to show up in places like that to stake their claim anytime soon - they're going to probably have to spend some of their own tax revenue to attract any flies, if the prospective market isn't big enough).

Of course, as Dennis Miller liked to say at the ends of his rants, "That's just my opinion, I could be wrong."

All the Best,
Joe Blow
Posted by Joe Blow (175 comments )
Reply Link Flag
The Story Says they Plan to Spend $15 Million ...
before all is said and done, but I think they meant that would be what was needed for a first phase of deployment and operation - the $6 - 8 million is probably just an alpha/beta initial capability in the areas of highest payoff - where people are mostly likely to want to use this capability beyond the existing capability provided by Starpukes, McDumbalds, etc. (and that would help keep the wee-wee contests with the corporate types to a dull(er) roar). I do agree that it will take far more than even the $15 million to be able to provide free access to every homeless guy's cardboard box around the city(just think about what kind of propagation/reflection headaches all those shopping carts full of aluminum cans are going to cause! :) )

However, remember that there is a paying option, and you can bet your sweet bippie that those customers will have a shiny (although, eventually, bird-poop covered ;) ) new access point within eyeshot of wherever they use their laptops. WiFi technology is much cheaper to field, even in the hundreds (thousands?) of nodes than cellular technology, especially if it's mesh-capable. Unlike cell phones, most current WiFi devices will not be used across nearly as much real estate (especially in mobile mode where the user is in control, e.g., walking, driving, etc.) - and anyone who does attempt that will soon be cleansed from the shallow end of the gene pool (but, why is it always _after_ they've reproduced?) by a cable car, trolley, bus, taxi, cyclist, mugger, etc. WiFi phones won't be a problem, as long as they are multi-mode and can take advantage of the existing cellular infrastructure (anyone know whether they default to WiFi when it's available? That will put additional strain on the network, if/when they start showing up in large numbers, but I hope for Earthlink's sake that it doesn't happen for a few years down the road/hill).

Other than proximity to Google's HQ, there is one very smart reason to start with San Francisco as an early nut to crack - if they can make this work there, it will be a relative breeze to make it work anywhere, even the glass and steel canyons of New York City. The continuing decline in the cost of the technology (and massive buying power) will also mean that cities they attempt this in elsewhere will be increasingly cheaper to do, from both hardware and labor costs (I'm guessing that propagation/usage models will be refined over time that will make each nth installment easier - but, still not necessarily easy).

It is curious that Cisco/Linksys has been silent about all of this commotion - maybe they're planning to buy/team with Yahoo/AOL/MSN to do the same thing in other places (or, maybe they're waiting for Google/Earthlink to prove/disprove the concept, and then do a me-too, or skip the mistake, in cities not already slated for this kind of capability. I wouldn't count out Microsloth leading its own team, either, with the obvious first deployment being Seattle (which may be third only to San Francisco and the sprawl of LA in geographic difficulty. Cities like Omaha are not going to be nearly as much of a challenge, now or in the future (and there won't be nearly as much money to make there, either, so don't expect a big rush of companies to show up in places like that to stake their claim anytime soon - they're going to probably have to spend some of their own tax revenue to attract any flies, if the prospective market isn't big enough).

Of course, as Dennis Miller liked to say at the ends of his rants, "That's just my opinion, I could be wrong."

All the Best,
Joe Blow
Posted by Joe Blow (175 comments )
Reply Link Flag
The Story Says they Plan to Spend $15 Million ...
before all is said and done, but I think they meant that would be what was needed for a first phase of deployment and operation - the $6 - 8 million is probably just an alpha/beta initial capability in the areas of highest payoff - where people are mostly likely to want to use this capability beyond the existing capability provided by Starpukes, McDumbalds, etc. (and that would help keep the oui-oui contests with the corporate types to a dull(er) roar). I do agree that it will take far more than even the $15 million to be able to provide free access to every homeless guy's cardboard box around the city(just think about what kind of propagation/reflection headaches all those shopping carts full of aluminum cans are going to cause! :) )

However, remember that there is a paying option, and you can bet your sweet bippie that those customers will have a shiny (although, eventually, bird-p..p covered ;) ) new access point within eyeshot of wherever they use their laptops. WiFi technology is much cheaper to field, even in the hundreds (thousands?) of nodes than cellular technology, especially if it's mesh-capable. Unlike cell phones, most current WiFi devices will not be used across nearly as much real estate (especially in mobile mode where the user is in control, e.g., walking, driving, etc.) - and anyone who does attempt that will soon be cleansed from the shallow end of the gene pool (but, why is it always _after_ they've reproduced?) by a cable car, trolley, bus, taxi, cyclist, mugger, etc. WiFi phones won't be a problem, as long as they are multi-mode and can take advantage of the existing cellular infrastructure (anyone know whether they default to WiFi when it's available? That will put additional strain on the network, if/when they start showing up in large numbers, but I hope for Earthlink's sake that it doesn't happen for a few years down the road/hill).

Other than proximity to Google's HQ, there is one very smart reason to start with San Francisco as an early nut to crack - if they can make this work there, it will be a relative breeze to make it work anywhere, even the glass and steel canyons of New York City. The continuing decline in the cost of the technology (and massive buying power) will also mean that cities they attempt this in elsewhere will be increasingly cheaper to do, from both hardware and labor costs (I'm guessing that propagation/usage models will be refined over time that will make each nth installment easier - but, still not necessarily easy).

It is curious that Cisco/Linksys has been silent about all of this commotion - maybe they're planning to buy/team with Yahoo/AOL/MSN to do the same thing in other places (or, maybe they're waiting for Google/Earthlink to prove/disprove the concept, and then do a me-too, or skip the mistake, in cities not already slated for this kind of capability. I wouldn't count out Microsloth leading its own team, either, with the obvious first deployment being Seattle (which may be third only to San Francisco and the sprawl of LA in geographic difficulty. Cities like Omaha are not going to be nearly as much of a challenge, now or in the future (and there won't be nearly as much money to make there, either, so don't expect a big rush of companies to show up in places like that to stake their claim anytime soon - they're going to probably have to spend some of their own tax revenue to attract any flies, if the prospective market isn't big enough).

Of course, as Dennis Miller liked to say at the ends of his rants, "That's just my opinion, I could be wrong."

All the Best,
Joe Blow
Posted by Joe Blow (175 comments )
Reply Link Flag
 

Join the conversation

Add your comment

The posting of advertisements, profanity, or personal attacks is prohibited. Click here to review our Terms of Use.

What's Hot

Discussions

Shared

RSS Feeds

Add headlines from CNET News to your homepage or feedreader.