Re: Freenet on a mesh network
Matthew Toseland <toad-EI5O+8PHWbJeeLb3ft/[email protected]> Wed, 6 Jul 2011 20:26:02 +0100
| Newsgroups | gmane.network.freenet.technical |
|---|---|
| Message-ID | <[email protected]> |
--===============1743673735== Content-Type: multipart/signed; boundary="nextPart5511731.iEYAlvFFN4"; protocol="application/pgp-signature"; micalg=pgp-sha256 Content-Transfer-Encoding: 7bit --nextPart5511731.iEYAlvFFN4 Content-Type: Text/Plain; charset="iso-8859-15" Content-Transfer-Encoding: quoted-printable On Tuesday 21 Jun 2011 21:07:42 Sheref Younan wrote: > Hi, > I'm wondering about the feasibility to use Freenet over a mesh network > independent of the internet. Where every Freenet node will be a node in t= he > mesh, nodes will communicate with neighbours via WiFi or other wireless > communication medium without the need of conventional Internet as a > communication medium. >=20 > This will have the advantage of offering a scalable and decentralized > network that will not need an existing Internet infrastructure and will > (among other things) resist block outs like the infamous one made in Egypt > during its revolution (January 2011) >=20 > any ideas, comments ? >=20 > Thanks, > Sheref Younan >=20 It's not a bad idea in principle. But there are many problems: 1. Freenet needs long links as well as short links. Any other scalable rout= ing protocol probably has the same issue, but Freenet in particular needs a= lot of short links and a few long links for routing to work. 2. Whatever you build, you'll need some way to exchange data between differ= ent towns or cities. One obvious possibility is to move data on a USB stick= or a hard disk. This won't be part of end-to-end routing for freenet (at l= east not until we have long term requests, in the distant future), but if y= ou are using freenet you could use binary blobs to move chosen content betw= een disconnected darknets in each city. 3. Generally wifi doesn't work all that well in pure ad hoc mode. Scaling i= s a problem above a few thousand nodes. Especially if everyone wants to use= the cheapest possible antennas at ground level (in theory directional link= s at rooftop level could scale a bit better). New technology (e.g. better M= IMO) may help. 4. Freenet cannot take advantage of various properties of radio that other = protocols might be able to. 5. Freenet tends to assume high uptime (nodes are switched on a large propo= rtion of the time), and in darknet mode (the safest) only talks to your fri= ends. If they are a long way away you'll need to route the packets there, a= nd it won't be very efficient, as they'll have to be repeated many times be= fore they even reach the next freenet-hop. And then multiply this by severa= l hops on freenet... 6. In principle the frequencies can be jammed and/or the nodes can be found. I'm not saying mesh networks and freenet don't combine for underground comm= unications... in principle it seems like a good idea ... Another key point: It may make sense to have automatic rendezvous when peop= le's phones are within range of each other, rather than fixed nodes. An "op= ennet" version of this is called Haggle, have a look at it. Being opennet, = they will talk to any node within range, so there is some worry over the fa= ct that your phone will be broadcasting your desire for the latest illegal = blog post... IMHO any underground network like that would probably have some parts be wi= fi, some of it very directional and hidden, some of it illegal satellite in= ternet hookups, and some parts sneakernet (either swapping disks for long r= ange transport, or possibly between friends for reasonable bandwidth but ve= ry high latency, or some sort of phone-rendezvous thing). Probably it would= have very high latency in parts, and Freenet can't do that just yet, but s= ome people (me) think it should eventually, or something like Freenet might. --nextPart5511731.iEYAlvFFN4 Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iEYEABEIAAYFAk4UtsoACgkQYUNbc3WUHYggnQCcDGWEFnwtTpj1jfn4ynKtTtNf AS4AoJet7hKaD4KKw2qCmOy3wtV+TiVw =uV0P -----END PGP SIGNATURE----- --nextPart5511731.iEYAlvFFN4-- --===============1743673735== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Tech mailing list [email protected] http://freenetproject.org/cgi-bin/mailman/listinfo/tech --===============1743673735==--