netBlazr Manifesto

A few weeks ago, Gordon Cook, publisher of the Cook Report on Internet and moderator of a well respected private email discussion entitled the Economics of IP Networks, asked me to explain to his list members what netBlazr was up to.  I got a bit impassioned and ended up included the background and motivations as well as our vision of how we will change the world.  So if you are interested, here’s what I wrote:

Thanks Gordon,

netBlazr is very much a pilot at the moment but we have great plans!
And let me apologize in advance if I run off at the keyboard, but you asked for it. 🙂

Motivation: I’m upset with the state of Internet connectivity in the US (and many other countries), but I’ve concluded there is nothing I can do in the political arena that will generate useful change, given existing vested interests (duopoly, FCC, Congress, etc.). However, I have observed that major political changes are possible when the commercial ground shifts. (E.g., the 1983 Bell divestiture agreement became possible once AT&T executives grew jealous of the booming computer industry and sought a way for AT&T to become a player in computers; or more recently, opening TV white spaces over broadcaster’s objections became feasible because of the commercial success of the computer and Internet industries, and 20 years of Wi-Fi market development). So the question arises, is there a way to radically change the commercial landscape in local access, even if it takes 10-15 years? I think yes.

As an aside, I’ve looked at quite a few community networks. Many are great, but too many are dependent on grants (that aren’t renewed) and on volunteers (who go off to grad school, etc.). To reach national or global scale, I concluded we needed to be a business, the more profitable the better (because that’s the way to scale rapidly). Ideally one finds a business model where the primary product or service can be free, funded by others who see benefit in the community that gathers around the primary product or service. We don’t have that nailed yet, but we have at least a “freemium” business model that is very workable and some ideas on how we might eventually go beyond freemium to become completely free for our primary users (that’s anyone seeking Internet connectivity).

Note that Internet connectivity is very low cost if you are in a data center located on the Internet backbone.  There’s lots of competition in backbone services. It’s also low cost if you are in a building that is “lit” by multiple competitive carriers (Cogent, Sidera, Abovenet, etc.), i.e. where multiple carriers own their own fibers into the building. However, if you are in a building with only monopoly or duopoly connections, even if it’s just 100 meters away, Internet connectivity can literally be 20x to 40x more expensive.

Networking in the home and the enterprise is also low cost – here it’s because you buy products in a competitive market, not services from a monopolist. Our problems in the US are local access and middle mile connectivity. Since fiber (& coax & copper pairs) requires rights-of-way, they will always have political overheads. And as a long time wireless aficionado (for example, search Google for “4G Tutorial” or “3G Tutorial”), it seemed reasonable to investigate wireless solutions. I did this at length in 2008-2009.

netBlazr uses point-to-point wireless links that are available at low cost because internally they leverage mass market Wi-Fi silicon. We use 5 GHz spectrum because higher frequencies are better for point-to-point links.  We go window-to-window to directly reach our users, and avoid their landlords (who frequently have inflated ideas of what common spaces are worth). We use directional antennas not to gain extra range but to reduce interference and thus leverage the substantial license-exempt spectrum available at 5 GHz (455 MHz in three bands in the US). So we’re tackling access issues first (middle mile later) as we can make many short high capacity hops at very low cost.

Of course, we’re interested in Free Space Optical (FSO) and 80 GHz radios and any other approach to high capacity point-to-point links, but today the only affordable wireless technology comes from leveraging mass market silicon. FSO devices cost many $K or tens of $K. At 5 GHz I can get a point-to-point link with close to 100 Mbps capacity for $135 (both ends), and Moore’s law is wind at our backs!

Topologically our network is a mesh, but it’s nothing like the “mesh networks” from Roofnet/Meraki or nearly two dozen other companies doing similar things. As we are using point-to-point links on different channels, our network is more like a bunch of routers connected by FastEthernet cables. We are early at this. Today the radios have to be aimed. Eventually (3-4 yrs?), we’ll get radios with electronic beam steering at mass market prices (the necessary silicon is in the pipeline).

So the issues for now are tweaking the business model to incent community formation and promote useful behaviors on the part of participants, and building the software platform to optimally support this rather different kind of wireless network. To that end, we have launched a network and are attaching customers, but so far only in one Boston neighborhood (Back Bay).

Thanks to those who’ve read this far. If you have questions, we’re trying to be completely open about what we’re doing and how our network is performing, so I’m happy to discuss anything at any level of detail.


  1. netBlazr, the Movement! | Bramka sms on February 14, 2011 at 5:37 am

    […] If you are interested, I’ve posted it on the netBlazr blog. […]

  2. Nikolay on April 8, 2011 at 5:09 pm

    the possible problem will be install costs and complexity… In standard Mesh it’s sort of self installed model. How do you handle that issue?

    Have you considered how to connect portable users around (who can use USB dongles for instance)?

    PS: what software is used for AAA and traffic shaoing purposes? FreeRadius or something similar?

    • Brough on April 11, 2011 at 3:42 pm

      Hi Nikolay, Like mesh networks, we are working towards complete self-installation. However, unlike mesh networks, our point-to-point links preserve available bandwidth, so we don’t suffer from 50% loss of capacity at each hop away from the Internet gateway. Also, unlike mesh networks (that typically have omni directional antennas) we are using directionality. So today we require some aiming. Eventually (with beam steering) that can also be automated. But yes, we are focused on businesses at first as they have more of a problem with the duopoly (and thus are willing to pay more for installation) and, with businesses, there is an available 3rd party installation force (IT outsourcing firms).

      We are currently focused just on fixed Internet access (primarily for small and medium businesses) in urban areas. We expect to branch to other services eventually, but not in the first 2+ years.

      We are currently building our service infrastructure and do leverage open source software whenever possible. We’ll have more to say about our service architecture when it’s a little more real. 🙂

  3. Nikolay on April 15, 2011 at 11:32 am

    I see. I’m interested in this topic, but considering probably to use 2 hope connectivity: 5G for p2p and business and 2.4G for public portable access. at that 2.4 APs can use dual band radio or be connected directly to the switch at 5G site. if you have any figures / estimations you may share let me know