Communities experimenting with mesh networks

http://technical.ly/2015/04/06/12-communities-experimenting-mesh-networks/

by carybeta • 11 years ago

34 14 11 years ago

I love this stuff.

I am taking a good Coursera course "Fog Networks and the Internet of Things" that covers edge networks. I am also planning on including a section "IoT and Edge Fog Networks" in my new book I am writing.

There is a lot that can be done at the edge (collecting and processing local data, sharing computational resources, etc.) without needing LTE/core or the cloud.

padobson• 11 years ago

Does anyone have any opinions on how this could grow to replace traditional ISPs? What's the difference between a mesh network and just putting a big WiMax antenna up in my neighborhood and giving everyone free internet access?

joepie91_• 11 years ago

The idea of a meshnet is that, as the name implies, it's a mesh. It usually involves intelligent rerouting over other peers if a previous route is no longer available.

If you were to put up a big WiMax antenna, you are the single point of failure - you could spy on the traffic, you could banish specific clients, and if you were to disappear, everybody in your neighbourhood would be without internet.

A mesh network allows none of these (in theory), because there is not one single point of control.

superuser2• 11 years ago

Mesh is a topology, not a belief system. A mesh network can still be administered by one entity who can monitor traffic, manage configuration, and restrict access to authorized users, but they would do so using a distributed system rather than the control panel of a single piece of equipment.

It's different from what you describe in that a typical WiMax antenna setup mirrors the star topology typical with network switches. Clients are connected to a piece of infrastructure, which is in turn connected to a larger piece of infrastructure, and so on up the tree.

In a mesh network, clients are connected to each other without using infrastructure as an intermediary. Every node relays traffic for its neighbors, and you reach distant points by hopping through many of your neighbors (or you don't, if they don't happen to currently be arranged in a way that facilitates that).

A good mesh network system has intelligent routing algorithms that are some combination of fault-tolerant, efficient at finding the shortest path, equitable in distributing load, and designed well in terms of incentive structure (so that people do not freeload by acting as clients but not relays).

If you are using a mesh network to reach the internet, you are still dependent on the nodes that route between the mesh network and an actual ISP, which presumably pay for the privilege and want to recoup costs somehow. However, it can be a decent way for a community to roll its own last-mile infrastructure, provided it can purchase wholesale transit somewhere within the mesh, and drive enough participation that there is sufficient capacity between you and the backhaul most of the time.

The more interesting uses are for meshes that are not (mostly) about Internet access. There is experimental work on protocols for messaging and durable document storage/publication (think Wikipedia but without web servers) that work well on meshes, providing pretty good internet-like capabilities to areas where access to the actual Internet is impractical. (Some of this is embedded in OLPC/XO).

Kalium4• 11 years ago

Mesh networks are suitable when all the computers you need to talk to are in a geographically contiguous area. As soon as you start needing to cross oceans or mountain ranges, you're likely to run off the edge of a mesh.

So I don't anticipate meshes replacing traditional ISPs any time soon.

chatmasta3• 11 years ago

Pure mesh networks will not take over ISP's. But what about a hybrid solution? Fat fiber pipe feeds wireless mesh. Run fiber from nearest datacenter to middle of town, build mesh network out from there.

Crowdfunding SEC rules + neutrality ISP rules could very well open up an opportunity for crowdfunding municipal ISP's. Communities run their own mesh and rent fiber from an upstream transit probider.

java-man2• 11 years ago

What I think might contribute to success of mesh networks as ISP alternatives is incorporation of micropayments for bandwidth/storage into the system.

There should be an incentive for individuals to participate. A thread of catastrophic communication system failure (as in case of a natural disaster) is not enough to stimulate the deployment. But the promise of income is.

Would you be inclined to install a mesh node if it gave you a small income stream?

Cloudy• 11 years ago

How would the income system work? (..cryptocoins?)

What goods/services can people buy with your "income stream" system?

java-man2• 11 years ago

Storage, bandwith, voice.

The problem, I think, is distributed accounting - in the presence of untrusted and malicious nodes. Micropayment transactions need to work faster than existing systems like BTC, to be able to support fast call set up, for example.

Communities experimenting with mesh networks

Related Stories