7 Comments

Summary:

Marty Cooper, one of the pioneers of cellphone business is not the retiring sort, and speaks his mind. At Broadband WorldWide Conference in Madrid, Spain he said that the concept of free wireless in the cities supported by advertising is a disruptive idea, but warned that […]

Marty Cooper, one of the pioneers of cellphone business is not the retiring sort, and speaks his mind. At Broadband WorldWide Conference in Madrid, Spain he said that the concept of free wireless in the cities supported by advertising is a disruptive idea, but warned that building wireless infrastructure in cities like San Francisco is not going to be either cheap or easy. He sees fatal flaws in Google’s plan to blanket the city with WiFi access points that would provide 95% outdoor and 90% indoor coverage.

We need to be careful here to separate the business and technology issues. I don’t believe it is possible to ‘blanket’ a place like San Francisco with Wireless LAN, and certainly not with the 300 access points per square mile Google is suggesting. I calculate they will need more like 3000 AP’s per square mile. They are clearly a great company, but I suggest they need a few lessons in wireless propagation.

He is not alone. Others are pegging the total cost of Google WiFi in San Francisco at about $250 million. The irony of his comments is that Intel has partnered with his company Arraycomm, for building smarter WiMAX antennas. At the same time, Intel is highly vested in WiFi and Muni-Wireless. But as I said, Marty is not shy to speak his mind.

Read More @ EE Times.

  1. I thought it was just me, but I don’t see why anyone would use wifi for WAN coverage with Wimax just around the corner. I could see Google acquiring GoRemote for wifi coverage in hotspots and related technology.

    Share
  2. Jesse Kopelman Friday, October 7, 2005

    The one mistake that people make with WiFi, especially at 2.4 GHz, is to worry only about signal strength and not about contention. The added contention you get from going from 300 to 3000 access points is going to eat up any speed gains you would expect to get from having the higher signal strength. Understanding this sort of diminishing returns is the art of large scale WiFi design. It is also why WiMax is far better suited for this type of application. That is a bit of a moot point for municipal networks until WiMax comes standard in lots of devices, however.

    Share
  3. Wi-Max is probably the best bet to cover the city even though Wi-Fi is cheaper.

    I have lots of experience in wi-fi (no kidding) and the monopoly I used to work for decided against free wi-fi only because of the sustainability costs. Do you know how much it costs to maintain those POP’s…per year. Google has no infrastructure experience..and their press release shows that.

    Share
  4. as much as free wifi is cool, i wonder if this will attract all the wifi thieves to san francisco?

    And regarding access to the SF pre-proposals:

    excuse regarding avilability of the proposals is not valid for a many reasons. the city continues to work behind the scenes on this issue and is trying to push through a solution without adequate community input.:

    “Chris Vein, who heads San Francisco’s technology office, declined to give many details about the proposals, citing the need to go over the documents first to excise any company secrets. The proposals, he said, range from very detailed to mere overviews”

    1. they got all the responses by Friday – so they could have been removing any secret information as proposals were received.
    2. After Friday 5pm, they could have removed the secrets over the weekend.
    3. they have not posted any of the proposals on the techconnect website for public review.
    4. in the Pre-RFI/RFC public meeting the city clearly stated that they could not really protect secret information since proposals would need to be public – so companies knew going in that they probably should not put that information in their submittals.
    5. this approach was reitereated in the written guidelines provided by the city.
    6. the city has been asked for a date as to when the proposals will be posted online and has not responded.
    7. the city has been asked to make sure that all proposals are in electronic form rather than paper only to allow posting on the internet – they have not responded.

    http://www.sfgov.org/site/tech_connect_page.asp?id=33899

    Share
  5. San Francisco isn’t Taipei, and the Taipei project is still getting going but … the raw statistics of the Nortel project in Taiwan suggest:
    10,000 access points in 90% of the city’s 270 square kilometers. So, it we do crude arithmetic on this, thus explicitly we ignore all topological and atmospheric and demographic (and other) factors: the Nortel/Taipei project has ~105 access points per square mile.
    Nortel execs (who clearly are not an unbiased source) say the project is going well; they also aver they’re using hardened APs, rather than consumer-grade gear. HOWEVER: to the points made by Marty and by your posters — it’s clearly time to bring in some field data on how well some large-area, urban and suburban installations are working.

    Share
  6. I love these carrier-trained guys, who forget that five 9’s, customer care, and the rest of it are out the window for networks like Google is suggesting (regardless of what the RFP might say). It might take a thousand access points instead of 300 but that’s small peanuts for Google if they pull this off. It’s also small peanuts for Hearst’s SF Chronicle who is the biggest loser in this scenario. Craig might have hurt them, but Google is out to kill them.

    Share
  7. Interesting post. Made me think a bit about the network and how Google will run it. The “network operatorâ€?(Google it seems in this case) will need to contend with ongoing infrastructure issues such as capacity, coverage and support for burgeoning apps like Voice over WiFi. But, is this within the realm of Google’s core competence? Google’s primary businesses have been information optimization and access (whether information in the form of location information (aka Google Maps), products & services, news, etc…). Is running a network related to this? Certainly being able to understand where and when end-users eyeballs are as well as whom the eyeballs belong to at any given time is extremely valuable to Google—it takes Keyword advertising to a completely new level.

    Marty is right on with respect to both potential coverage and capacity issues – blanketing San Fran will take much more than Google’s suggested 300 access points per square mile.

    Also, let’s remember that “free is never freeâ€? –someone is funding it, shareholders/ads….and the goal of “freeâ€? or reduced fees for a product or service, is generally intended to drive the sales of a different product or service (give away razors to sell the blades…). Google has built a very successful ad revenue stream—the “holy grailâ€? for those offering advertising is being able to deliver eyeballs and click-throughs for an intended, targeted audience to an advertiser and then provide data back to the advertiser that its key audience is consistently being ‘hit.’ Might “freeâ€? Wi-Fi do it? Further, if they can subsidize the “freeâ€? Wi-Fi by also reselling wholesale Wi-Fi network access to other service providers, they mitigate the risk of the investment. That seems to me how/why this makes sense for Google.

    Share

Comments have been disabled for this post