Buy Broadband From Your Neighbor

infractor writes "Wired has an article about a wireless project delivering free broadband to a rural community. Using Linux based devices called meshboxes from Locustworld, they've created a local mesh network. More detail in this article. With Wi-Fi friendly ISPs talking about micro-ISP deals for wireless sharers this could be the accelerator UK broadband has been waiting for." Last year we mentioned the MeshAP-05, a bootable CD which "turns a single board computer or laptop into a mesh node and access point," since updated to MeshAP-06. Update: 02/13 19:52 GMT by T : I see from comments that -08 is actually the current version of MeshAP, with -09 soon. Thanks.

Anybody else think...

[Anonvmous Coward]

.. one day the internet will be rivaled by a community born network? The pieces are almost in place, networking's cheap and easy, peer to peer, desire to do it, etc.

A year or two ago I couldn't imagine it, but I can today. Two of the apartment complexes I've lived in I had neighbors that would have been interested in networking their computers with mine. If wireless had come around sooner (price-wise I mean) we would likely have done it.

Okay, I'm not really on topic. It's just this article put an interesting image in my mind of what I'll be connecting to within the next 5 years.

Re:telco's

[AmigaAvenger]

Except most of the globe doesn't have civilization anywhere near it. Even rural states would only have very localized coverage, nothing close to 'global' (IE.. My closest neighbor is 20 miles away, doesn't really pay to set up the towers needed for this type of access)

In Soviet Russia...

[Hadur]

Please read before modding down automaticaly :) Anyways, this sounds like a great idea on paper. But, it seems to be relying on one thing: human goodness. Communism also sounds good on paper. In fact, it is utopia. But, it will NEVER work as it should because humans are greedy. In this case, what is stopping anyone from geting on Kazaa and using up all the available bandwidth? Well, there goes the high speeds for the other neighbors out the window. Yes, they could impliment some sort of bandwidth throttling, but where do you cut it at? You would need a speed at which is fast enough to make the technology and effort viable, but a speed slow enough to prevent misuse... A hard thing to decide. So, in conclusion, I agree that this is a very interesting new application of technology in theory, but in actual implimenttion, I see some serious design flaws.

Re:telco's

[CharlieO]

Errr - do you have any idea of the size of traffic carried by backhaul telcos?

Or the very very serious iron needed in switches and management systems to make sure it works?

And who gets to decide the routing priority in these networks?

Who gets to warrant the privacy of data? Telecoms companies are bound by some pretty strong laws to protect the privacy of the voice and data traffic they carry - home supported APs wont.

Re:Getting the broadband in the first place

[silentbozo]

You can't hook up your wireless network point to a 56k modem and share that about.

Why not? Apple's Airport base station has this capability. The meshboxes sold by LocustWorld (as mentioned in the article) are standard PCs adapted for use as low-power, low-heat, high-reliability base stations - I imagine that hacking the stack to route packets from a modem to the rest of the network would be trivial. Even better, forget hacking the meshbox - just set up a NAT with a dialup on the other end, and DHCP access to everybody else on the wireless end.

Sure, it'll be slow as hell. Maybe someone will cache commonly accessed stuff on a daily (or semi-daily) basis to reduce bandwidth load and access time. However 56k is more than enough for basic e-mail, and low-bandwidth web surfing. In the meantime, you build a wireless community that maybe, one day, will have enough users to pony up and put in a leased line, and retire that old 56k modem. :)

No, I don't think...

[Donut]

You are spending too much time on Kazaa. The Internet, and the majority of its casual users, hit a very small number of high traffic sites (Slashdot, google, CNN) that are sitting on very fat pipes. While these sites are distributed somewhat to different geographic locations, it is still very centralized, and not very peer-to-peer.

While it is can be argued that the end points of the small-time user part of the Net may become free from certain ISP based constraints, there will always be a need for Telcos and their fat pipes for a majority of the mainstream content on the web.

-Donut

ps. Before you grip about homogenous content being the death of freedom, reflect on how much more diverse the net is to the bygone days of the Big Three TV networks.

Re:telco's

[JamesSharman]

Actualy I think the original poster has a point. The suggestion isn't that this will happen over night but imagine the development process.

First you get clusters that are only linked via the established networked.

Then clusters of those get linked up with longer range networking techniques (Do I need to mention Pringles can antenas)

Eventualy you only need to cover the large geographic areas between population centers but thats where something like shortwave gets in.

Sure speed is proberbly going to be dependant on distance but the suggestion that one day we may do most of our global networking independant of the telcos sounds pretty good to me.

The only problem with the article and some thought

[zymano]

The article never covered how much distance their Wi-Fi signals travel. If anyone has some links to how far the signal goes, i would thank you. Wouldn't a better idea be to put microwave receiving dishes on everyone's house ? You could get satellite and ground link microwave reception.Didn't the gov release some of the frequency for the public. We need to take back the radio spectrum from the government who just sells to highest bidder . It's a pure ripoff if you ask me. The radio spectrum should be FREE . It belongs to the people and not Corporate america .

What's really needed here...

[Adeptus_Luminati]

... is:

1) METHOD TO DEAL WITH PER GIG COSTS PER MONTH:
To have software installed on each of these computers that are connected to the broadband access directly to monitor how much traffic they've sent and received that month.This should be simple enough to accomplish. I say this because if this thing really takes off, it won't be long before Telco's clue in and start charging per gig per month for direct broadband users. With such software the user willing to share his broadband connection to a comfortable threshold limit... say 50% of his 'free-bandwidth-before-he-has-to-pay-additional-ch arges-to-his-ISP-per-month'. This allows the wireless network to grow at a rate that is sustainable by the people willing to share their broadband access. It also encourages others who have direct broadband to share their connection with nearly ZERO risk of having to pay additional monthly charges. (this is sort of already done with Kazaa, where you can limit the upload speed, how many users can download from you etc.. only we'd need, max output per month, max speed per second, max users sharing service -- CONTROLS.)

2) DONATION/PAYMENT AUTHENTICATION PROTOCOL: Imagine a wireless user turns on his laptop in an area with multiple shared broadband connections, a dialog box comes up displaying a list of 10 different connections he can choose from. This list would be sortable by: available speed, cost per gig, max users, etc. The laptop wireless user then can click on the cheapest connection, or the one with the most available bandwidth (if he has deeper pockets), and start surfing the net. The donation authentication protocol would allow the laptop user to automagically transfer funds from his paypal (or-insert-future-online- digital-fund-transfer-systems-here) to the broad band service provider (the user sharing his DSL/cable modem), and thus we have created:
a) A cost per use wireless network
b) A method to allow for individual directly broadband connected individuals to have free internet access (their monthly fees would be paid by their wireless customers)

A WIN:WIN for everyone? I think so... even the telcos could benefit if they choose to start charging per gig.. that would just end up eventually defining more precisely the cost per meg/gig a wireless user would have to pay depending on the area he's in.

Re:Clarification

[nule.org]

This project seems really cool, and I want to run a node, except I have no desire to dedicate a piece of hardware to it. Is there any plan to release a package that can be installed on an existing workstation to provide the mesh functionality, yet still allow the user to use the workstation for other things?

I tried to find the answer to this myself but the downloads section of the site is a bit confusing.

Re:telco's

[Anonymous Coward]

>I'd guess that data transfer rates drop at >least linearly with distance from "static" >(non-ad-hoc) routing nodes.
I can believe that, but there is still
potential win in robustness. Currently it
just takes one idiot digging up fibre to isolate
an area completely. With a good mesh network,
traffic can fall back on the nearest static
node. Of course anyone hoping to download their
next year's worth of music is out of luck at
that point, but for many people there's a huge
practical difference between "it's gone very
slow but e-mail, DNS etc. are still getting
through" and "it's dead until further notice".

Re:Nice one guys - slashdot a group of hard up peo

[Pyrosz]

This is a very good point. Slashdot should implement some sort of policy about such things. The /. effect can be very detrimental to these small sites and can have them shut down as a result. Yet again the idea of having a /. cache of the site is valid, I'm sure they will not sue for copyright violations if /. is saving them a ton of money.

Oh well, I expect this post to be ignored like all the rest by the Slashdot editors. They are not very professional when it comes to these things, but here I am again preaching to the choir.

T1: Landlords as ISPs

[axxackall]

DSL? Cable? Again? Oh no!

I have much better idea to propose to landlords of big appartment buidlings:

Make a deal with some good ISP, get a T1 from them to the building, put Linux server there in the building, and sell the connection to your tenants.

Most of modern building have enough of C5 phone cables, so the access media should not be a problem. Otherwise - wireless.

Tenants can have even own web servers. One option: if the landlord rents a class C subnet. Another option: use that Linux router as a frontend (NAT or proxy - your choice).

I hate DHCP of most of DSL and cable providers. And it's hard to find good ISP with static address, high speed and low price. I think it's realistic to calculate the business model in a way to share that T1 for $40 per tenant monthly.

Re:Accountability IS important

[dvdeug]

Or put it this way, would you be happy to hand your credit card details

Whatever happened to encryption? Any one on a local network can packet sniff your information anyway, so handle it properly and no one can read it.

I can show you the agreements I have to sign that make me legally accountable for protecting any information I am privy to in the use of those licenses.

I'm sure Martha Stewart and the heads of Enron had to sign papers saying they would not embezzle, cheat and scam their way to fortune. It didn't stop them.

Re:In Soviet Russia...

[inKubus]

Yes, but in a large enough mesh network, chances are your neighbor would have the file you need; no need to go out on the internet for it. The idea is a network fabric, hopefully with many outgoing routed connections to it as possible. It's not like most people with 1.5 MB cable connections use all that bandwidth all the time also. Might as well share. Someone has to just do it.

Re:Groovy.

[mapMonkey]

Yet, you still give them your money? I'm confused.

Bandwidth vs. Latency

[dprice]

So what is the performance of a mesh network built out of 802.11 nodes? Many people would say 11 Mbps to 54 Mbps minus the usual overhead depending on the type of 802.11 being used, but raw bandwidth is only a piece of the overall performance.

I would think that latency would be the main limiter of a mesh network. The nodes would have to be placed relatively close together if built with off-the-shelf 802.11 equipment, so it would take quite a few hops to traverse any long distance. Each node would have to analyze and route the traffic which adds further latency.

I also wonder what the scalability of a such mesh network is. As the mesh grows to a large number of nodes, I imagine that congested hot spots will develop which will add latency as traffic waits to be processed or has to route around the congestion. I wouldn't be surprised if packets could take minutes to get across country if only a mesh network is used.

For a small number of nodes, the mesh probably provides a reasonable solution for small networks and for providing the "last mile" from a conventional wired internet connection. For latency tolerant applications like email, a larger mesh might be acceptable (anyone remember Fidonet?). I have my doubts that a large mesh could be used as an equivalent replacement for a wired internet.