Hacker News Reader

Comment by michaelt | Hacker News Reader

I'm no expert on mesh networks so maybe someone can tell me - how would a mesh network connect Los Angeles and Las Vegas? Or San Francisco to Salt Lake City?

Given the populations on both ends, I assume you'd need more bandwidth than you could get from wifi routers with pringles can antennas.

I assume the companies who own long distance fiber have to do what the government says, as they need government permission to dig up the streets and install more fiber meaning they can't afford to upset politicians. So you can encrypt the traffic, but if the word comes down to cut the cable the cable gets cut.

View on HN

Yeah, bandwidth is the big thing people overlook. The land-based backbones across the country is already made of bundles of 10Gbit strands. Can you imagine going from that, to all of the nation's traffic over a couple 54Mbps streams...

And these backbones are all in the hands of companies and governments very much against these proposals, or anything offering freedom. Don't discount the demise of Net Neutrality either, they'll filter out all these tunnels or just store them in Bluffdale for future decryption and prosecution. Unless of course, you are one of the good guys with private backbone fibers… (see my HN profile for those fibers and open hardware optical switches).

Even if traffic is drastically slower, on the order of a dialup modem, there is still be massive value in free internet for the poor and the developing world. Text is a powerful medium.

Let's figure we can get 10 WiFi links at half of full speed across the country, 10 x 30Mbps. That's 300Mbps, or roughly 1 bit per second per person. About 10 letters per minute.

Good luck. Even the very first modems offered 300 bits per second, so you would need 3,000 simultaneous non-colliding WiFi links across the country to have enough bandwidth for everyone to enjoy the speeds of the very first modems - a cool 40 letters per second.

"About 10 letters per minute."

This is a OpenLibreNet telegram. Full stop.

Assuming everyone is using it all the time, and there's no meaningful caching that can be done.

A mesh network is not necessarily the same thing as a wireless mesh network.

But that is how this mesh is supposed to work according to point 3.

It being a mesh protocol would supposedly mean that cut cables would get routed around fairly automatically. You'd have to cut all the cables as there'd always be a financial incentive for someone to carry that data to its destination.

And there is more than just wifi to pick up the slack in an emergency. Free space optical links, satellite uplinks are all possible mediums. I'm not arguing that this would work well, but I think this is the idea behind it all.

Not too long ago, the internet measured top speeds in tens of kBps. So first off, give it time. Engineers are good at solving problems like this.

Furthermore, mesh networking is a fundamentally different architecture solving a different problem than the internet. The current internet works great for high-bandwidth applications like commercial video streaming, but is terrible in terms of protecting itself from concerted attack by powerful actors. Mesh networks solve the single-point-of-failure problem.

http://gizmodo.com/5912383/how-to-destroy-the-internet

Not too far into their paper, they posit "tunnel nodes" that connect isolated mesh clusters through the normal Internet.

As the network grows, they imagine that tunnelling through the Internet will become rarer.

I don't think we have the technology today make a "global mesh network". We need new Wi-Fi technologies and products to appear, and I think they will over the next 10 years or so. We need to get the Wi-Fi Alliance to create new Wi-Fi standards for longer ranges, too, which would make this easier. It would also help commercial companies, especially if they want to disrupt wireless telcos (such as Google and Microsoft, who have been trying to work on this, too).

But until we get to that point, we can build the technology, and have it more localized in villages, towns, or even big cities. We can use the latest Wi-Fi routers, or mobile phones, to form these meshnets now. Then perhaps we'll get Wi-fi balloons like what Google is building in Africa, although I'm not sure how resilient those would be against a government, but perhaps there will be so many of them that shutting them down would be much like shutting down the Internet.

As for intercontinental meshnets, I think DIY sattellites will become more popular in the future, and could help with this, too.

Think low-cost drones [1,2], not satellites. I can imagine a constellation of drones, linked by modulated laser/led/radio beams beamed though the air. If using light/IR, a challenge would be keeping the beam pointing at the photo-diode (optical phased array?). The drones could link to the ground using either optical or radio. An awful lot of this work is already underway in the context of satellites.

[1] http://oilprice.com/Latest-Energy-News/World-News/Solar-Powe... [2] http://www.wired.com/dangerroom/2013/08/new-9-hour-solar-pow...

They forget about wear and tear. The bandwidth is limited, as you should measure it in low-power high-bits/cubic-meter[1]. The higher the drone, the greater the distance therefore the lower the bandwith per cubic meter. A fast mesh network needs optical or should have ultra low power butterfly like drones with many gigabits, say one butterfly per human or per cubic meter.

[1] http://thor.cs.ucsb.edu/~ravenben/papers/coreos/wei93.pdf

A related question: Does the proposed protocol allow the mesh network to be run on top of the existing Internet?

For example, it may be done with the satellites, which are sent by the community. Kickstarter already helped similar projects.

Reply on news.ycombinator.com