Mesh networking: another disruptive technology

Jon Hannibal Stokes has a thoughtful piece on ArsTechnica about Meraki Networks, a start-up which is commercialising networking technology that emerged from the MIT Roofnet project.

In a nutshell, MIT’s Roofnet allows people in and around Central Square in Cambridge to gang together their wireless access points into a kind of wireless cloud that anyone with a WiFi device can access if they’re in range. There are some specifics I’m leaving out—you have to use a particular model of router, and you have to sign up for the program—but you get the general idea.

There are two ways to participate in Roofnet as a wireless access provider: as a node on the mesh, or as a gateway. If you participate as a node, then all you do is put the right model of wireless router running the right software in your window and turn it on. The router connects to other, nearby wireless routers, and it routes packets for the network and acts as an access point for end users. Of course, there have to be wired connections providing Internet connectivity somewhere in the mesh, and that’s where the gateways come in. If you participate in Roofnet as a gateway, then you’re sharing your own personal cable or DSL bandwidth with the rest of the network.

Meraki Networks plans to commercialize this mesh network model by offering a small, cheap ($50) wireless router, the Meraki Mini, that comes pre-loaded with the mesh network software. You can use the Mini to launch your own wireless network by just plugging it into your own broadband connection. The Mini’s software lets you do traffic monitoring and shaping, branding, and billing, so that you’re essentially reselling the bandwidth of a company like Comcast or AT&T. (Yeah, the telcos are gonna love that idea, but more on that in a moment.) Other users with Minis can connect to your router and extend the network outwards, choosing to participate as nodes or as gateways. With enough of these devices, you could cover a whole apartment building, or a whole block, with wireless… that is, if they don’t step on each other.

As Stokes points out, there are lots of interesting potential problems here. Some are technical — e.g. what happens when the mesh becomes very dense and interference starts to become a real problem? But it doesn’t take a rocket scientist to spot the other, more intractable, problems.

As for the legal challenges, everyone from the federal government to the RIAA to your broadband provider are going to want a piece of you the moment you hang out an ISP shingle and start billing customers. Will you be obliged to comply with CALEA if you choose to route VoIP traffic? Will your (quasi?) official status as an “ISP” grant you immunity to RIAA lawsuits while making you the target of subpoenas instead? Is Comcast really going to sit still while the number of wired Internet connections in an apartment block drops by half or more, with their remaining customers acting as competition by reselling Comcast’s own bandwidth to former customers?

Of course, not all ISPs are like Comcast and Verizon, which forbid sharing your wireless connection with others. Speakeasy, for instance, actively encourages their users to share their connection with the public. I think ISPs could get creative and ask for a cut of the proceeds that users get from reselling bandwidth, in effect making their end users authorized bandwidth resellers. But that idea makes sense, and when it comes to anything that smells of “P2P” and “grassroots,” rationality rarely prevails in the boardroom.

There’s an extra angle to this in the UK, in that I think that it’s actually illegal under the provisions of one of the Communications Acts for an ordinary person to sell bandwidth. (I can freely share my wireless network with my neighbour, but I couldn’t sell him airtime.)

Of course the guys who set up Meraki know all this, which is where the interesting bit comes in. The NYT reports that Google and Sequoia Capital have invested in the company.

Stay tuned.