Home » Internet

Google Broadband isn’t practical at a national scale

By 11 February 2010 15 Comments

Google is planning to deploy an experimental high speed broadband testbed with 50,000 to 500,000 people and the blogosphere is already going into hype overdrive.  The true goal of the project is not to build a new national broadband network but to “rabble-rouse” the existing broadband providers by demonstrate how things ought to be done in the mind of Google.  For example; Google’s demonstration broadband network is going to offer gigabit speeds at “competitive prices”, only provide generic IP service and not Telephony or Television service, and offer open access wholesale to other ISPs.  In other words, Google Broadband will be Google’s Utopian vision of how all broadband providers should be and what they’ve been advocating in DC all these years.

Of course the reality is that no broadband provider can all of these things or even most of them.  Even the best examples of open access wholesale in Europe don’t demand mandate open access on the pure fiber networks because no network operator is going to spend billions of dollars on a Fiber to the Home (FTTH) network that they have to open up to their competitors at regulated wholesale prices.  Furthermore, even the municipal fiber networks have to rely on good TV and business class service uptake to break even or face financial disaster.  Dr. Timothy Nulty who heads one of the most successful municipal fiber networks in Burlington Vermont stated that the open access wholesale model was “a recipe for financial failure“.

The other big difference in Google Broadband is that they are going to cherry pick communities that will be cheap to deploy which is not something you could ever replicate on a national level where deployment levels have to be above 90%.  Google is most likely going to deploy at the cheapest locations where fiber runs are short and where aerial cabling will be permitted.  Underground fiber cabling can cost nearly 7 times more according to this study performed in San Francisco.  The CTO of Qwest has stated that they won’t deploy fiber to the home because 75% to 80% of their cable plant is underground whereas Verizon is the opposite, and that makes a huge difference in the cost of fiber deployment.  So Google Broadband might scale wonderfully when they can cherry pick 20 to 200 thousand homes (assuming 2.5 people per home), but it’s a whole different ball game when we’re trying to reach 100 million homes.

Then there’s the matter of actual speed versus advertised speed.  Can Google actually deliver anything close to gigabit speeds to all of their users during peak usage times?  It’s highly unlikely but that hasn’t stopped journalists from declaring Verizon FiOS (an actual production service you could buy and not just some vaporware service) as “also-rans“.  But what most people don’t understand is that Verizon FiOS is already a 2.4 Gigabit GPON infrastructure that is already testing 10 gigabit GX-PON technology.  But for now, Verizon is offering speeds that they can actually deliver or exceed so that when a customer buys 20 Mbps service, they actually get 20 Mbps of actual data throughput.  If Verizon wanted to be more like Japanese broadband providers which typical deliver less than 1/3 of the OECD advertised speed data, they could easily claim to offer “100 Mbps” or “1000 Mbps” broadband service so long as they didn’t care to maintain their current level of customer satisfaction.

These are very complex issues that few understand so the typical reaction if Google Broadband ever takes off is “why can’t my ISP be just like Google”, and that’s probably what Google is counting on.  But a truly useful discussion on Broadband should be based more on fact and less on hype.


  • Michael Turk said:

    The other big difference in Google Broadband is that they are going to cherry pick communities that will be cheap to deploy which is not something you could ever replicate on a national level where deployment levels have to be above 90%.

    That’s the part of their plan that tells me they recognize this is not sustainable on a broad scale. They specifically note “local factors that will impact the efficiency and speed of our deployment, such as the level of community support, local resources, weather conditions, approved construction methods and local regulatory issues.”

    In other words, if we can find cheap labor, no opposition, ideal weather, and lax construction rules, we’ll go for it.

    If they wanted to make this testbed realistic, they would pick the place with crappy weather, onerous local regulations, crooked politicians, an angry community, and multiple established providers. That way it would look like most of the country.

  • jhn said:

    Nothing you write is a dig on their plan. They’re looking to build a couple of “testbeds” where they see what kind of use people put a 1 gb network to. They’re trying to demonstrate that a network run as they say a network should be can work. If the network is sustainable as an open access network, we’ll find out. If not, we’ll find out that, too. (Personally, while some competition is better than none, shared-infrastructure competition will never get you to the point where you can have dozens of ISPs, even if it does “work.”)

    I don’t think they’ve said they think that fiber is viable for every last rural household, or even politically doable, given the amount of digging required. But, if fiber is doable in some communities, why not have it in those communities? Gigabit speeds are advertised and delivered in some markets–dense places like Hong Kong. If it turns out that you can have affordable backhaul and gigabit last mile in Manhattan, isn’t that a good thing?

    It’s nice that the backlash has changed from being “they want to start subsidized municipal networks and destroy competition.” You might think that Google shouldn’t be pushing what you see as bad policies on the FCC. But here, Google is at least putting its money where its mouth is. They’re not looking for special favors or handouts.

  • George Ou (author) said:


    The problem is that Google is looking to deploy in the cheapest 1% (or less) of the nation, but it probably costs 10 times more in the remaining 99% of the nation. But Google and the other groups like Free Press will try to claim that if Google can do this, then why can’t the rest of the country. That is an absurd argument because what you can do in the cheapest 1% has no applicability to the rest of the nation.

  • Sami Samhuri said:

    George, progress has to start somewhere. You can’t hold everyone back and move in lock step in a country the size of the US. How could this be anything but good?

    > Can Google actually deliver anything close to gigabit speeds to all of their users during peak usage times? It’s highly unlikely […]

    Do you have anything to back that up? We know they have been purchasing dark fibre since at least 2005 or so. They have data centres around the US. We don’t know if they can do it, but it seems you don’t want them to even try. Why is that?

  • Aaron Huslage said:

    Google’s trial is absolutely feasible. You’re basing your ideas on existing business models for ISPs, which are not the *only* way to do things. Google is being a disrupter of the first order here. They are bringing and testing new ways of building and charging for things that don’t come with legacy telco baggage. Just because we’ve always done it one way that means it has to be done that way forever…that way be dinosaurs.

  • George Ou (author) said:

    To Aaron and Sami,

    No one is holding anything back. I have no problem with Google trying and I’ll be the first to buy the service if it’s fast, competitively priced, reliable, and available to me. But you two seems to have missed my points. The lessons learned in the Google experiment apply to 1% of the nation where Google will cherry pick cheapest places to deploy, but it does not apply to the remaining 99% of the nation because the costs will be several times greater in those instances.

    Aaron, judging by your comments, I really doubt you’ve actually looked at what it takes to build a network. You seem to be saying that all the existing ISPs are stupid and they’re doing everything wrong. But if that’s the case, Google should be able to come in and build a nationwide broadband network and murder the existing “dinosaurs” and make a mint while doing it. But why is Google stopping at 50,000 to 500,000 people when there are 300,000,000 people in the country?

  • Todd said:

    I don’t think that existing ISPs are stupid nor are they doing anything wrong *for their business model*. Existing ISPs are only motivated to improve access and performance of their networks when there is direct competition. Without competition it makes no sense at all for them to do anything other then keep cashing monthly access checks.

    Google, however, has a very different business model. Their model is basically tied to the performance of the internet. They need everyone to have fast hi-bandwidth access to the internet. In fact, they need it so much they are willing to fund it.

    This is what makes Google such a game changer. And also it is what aligns them so well (for now) with the needs of the market. They need what we need. A super fast, reliable network, available to everyone, everywhere. You can see this motivation in a lot of their recent announcements, like their DNS offering, and Chrome, Android etc. It is motivated by a desire to get more internet into more hands.

    This is a pretty good for everyone ( expect ISPs ) for now, although I do foresee a day when this is no longer the case.


  • George Ou (author) said:

    Todd says: “They need everyone to have fast hi-bandwidth access to the internet. In fact, they need it so much they are willing to fund it.”

    Nonsense Todd. The focus on the access network is misplaced. 95% of the nation has access to 1 wired broadband provider and probably 86% have access to 2 or more wired providers. Most people can spend a few more dollars a month to double their Internet speed but they don’t and I don’t blame them because there’s no content that needs it.

    If Google was really serious about spurring on faster broadband, why are they delivering over-compressed 1080P video on YouTube at 3.75 Mbps at best? Why isn’t Google offering at least 6, 10, 20, or 40 Mbps Blu-Ray quality content on YouTube so that more people will have a reason to buy faster broadband services?

    Pushing for faster broadband when the content isn’t there is simply putting the horse before the cart and that’s the trap the Japanese broadband market has fallen into. In fact, content hosting bandwidth in Japan is several times more expensive than the US which is causing their content to run at much lower bitrates than US based YouTube even though they have twice the measured broadband capacity. What they have is an even less balanced Internet where users have fast pipes but all the content is narrower than the US based content.

  • Todd said:

    You ask ” Why isn’t Google offering at least 6, 10, 20, or 40 Mbps Blu-Ray quality content on YouTube so that more people will have a reason to buy faster broadband services?”

    Why on earth would they do that? Of course no one would buy it. That also doesn’t solve the problem fast enough. nor does it solve the problem in a way that is beneficial to Google. At best it would drive money into the hands of Googles competitors.

    They do not need to do that. All the need to do is get people to use their stuff, weather its gmail, docs, youtube, App Engine, Android, Buzz, and maybe now their broadband network. The magic of Adwords and Adsense does the rest.

    You say that google is “putting the cart before the horse”, byt trying to create a network for content that doesn’t exist yet. This is exactly what market creators do. They build it and people come. There will be no content for it until there is a network to support it.

    There was no market for 100,000 mobile applications until Apple built a platform to deliver it.


  • George Ou (author) said:

    Todd says: “They do not need to do that. All the need to do is get people to use their stuff, weather its gmail, docs, youtube, App Engine, Android, Buzz, and maybe now their broadband network. The magic of Adwords and Adsense does the rest.”

    I asked you a simple question why Google is not offering people the kinds of applications like higher quality YouTube that would get people to spend a little more money on faster broadband and this is the best you could answer?

    Perhaps you should apply for a job at Google PR with that kind of spin and rhetoric.

  • Digital Society » Blog Archive » Podcast: Google: The ISP said:

    […] Google Broadband isn’t practical at a national scale – George Ou […]

  • Todd said:

    I assure you I am not trying to spin for Google. I am actually quite concerned about a Google dominated future. I think there are serious privacy issues that have yet to come home to roost. Any company that has to tell you “don’t worry, we are not evil” is very very scary.

    What I am trying to do is understand what they are up to. Their business model frees them from the constraints that face almost all their competitors, and as a result they are blowing up markets left and right. I am not sure why the broadband market will be any different.

    For now I am reasonable happy with the bargain I have struck with Google. I give them my data and they let me use their stuff for free. That will almost certainly change in the future.

  • Rudolf van der Berg said:

    Reggefiber in The Netherlands is an open network by its own choice. It has several operators operating over its network. It had this already before it was regulated and it now that it is regulated this is still the case and it isn’t opposing this regulation as it is good for business. The businesscase for Regge isn’t about services. the businesscase is a pure real estate play and in real estate you don’t care about the business of your tenants as long as they pay the rent. Oh btw one of its service providers is offering 200mbit/s symmetrical internet.

    KPN the Dutch incumbent has publicly stated it is happy to run an open access DSL network and has stated that opposing the opening of the network was a bad business decision.

    The same goes for the Stokab network in Stockholm, which is an open access network.

    Also have a look at Free.fr in France, who are in favor of open networks and who offer a highly competitive network at 100mbit/s and yes they do deliver on their promises.

    Oh well.. you wanted a discussion based on facts… don’t look to Europe, they will mess up your opinions. :-)

  • George Ou (author) said:

    Todd, there’s no magical business model that’s going to all of a sudden lower the brutal costs of running a broadband network.

  • Cayble said:

    future where they will need fast hi-speed band width. That much is bluntly clear.

    I am constantly amazed at the general lack of reporting on what seems to be empirically obvious directions all the top tier IT leader companies are pushing towards. That direction is clearly a future where personal hardware ownership is of a significantly diminished importance, personal software ownership is nonexistent and actual possession of said software is likewise non existent.

    The entire IT industry, well at least the eggheads who run the large companies, have borne witness to the debacle that Windows XP has proved to be for Microsoft, and of course many lesser software companies have felt similar pinches from lack of interest in new offerings when the years old offerings are good enough, and the new offerings are far from free. Nobody wants to be in the position Microsoft finds themselves in with Windows XP. Its also really hurting the hardware side of computing. Its just great if you have a perfectly fine OS that’s years old and you have no interest in upgrading it, particularly if the hardware is still good enough to get by.

    They all want a world where the hardware is smaller, lighter, cheaper and the software is rented, the storage is largely rented and the bandwidth is never pinched due to the incredible amount of traffic that will al generate. Neither Google or Apple want to try anymore to defeat Windows head on. Its an unwinnable fight the way the arena is currently set up. Now they want to change the arena, and even Microsoft sees the value in that. Microsoft, just like the rest, would love to see a new paradigm where you rent ALL the software, including whatever will be passing for an OS in ten years, and they can just send through free upgrades at their leisure all included in the monthly rental costs. Tiny SSD, or similar drives will be all that’s needed as they will want you to store all your content and data in the cloud someplace…at another cost.

    So ya, Google is looking to the future, they want to produce operating systems that are largely browsers, using cloud based apps and storage, and they cannot risk being ready to roll out the program in any serious fashion if it would lead to clogging up a less then speedy system.

    So ya, make no mistake, Google dosnt NEED it today, they NEED to get it started today so it will be available tomorrow.