The IT Crowd's internet in a box gets $240k of crowdcash for a cause Back in February we reported on a crew called Outernet and its plan to float a network of tiny satellites to deliver internet access around the world and in the process “bypass censorship, ensure privacy, and offer a universally-accessible information service at no cost to global citizens.” Outernet's made some progress …
Why the Hell the link purportedly leading to a previous story about Outernet actually sends us to a piece on an RBS and Natwest titsup from 2012? Is it some sort of practical joke non-Brits aren't supposed to get?
The original Outernet story is here, by the way.
Really, there's no need to rush stuff this badly, I doubt anyone would have bothered if it took another couple hours to hear about this.
...“bypass censorship, ensure privacy, and offer a universally-accessible information service at no cost to global citizens.”
How, exactly is that going to work then? If your box is preloaded and requires a data connection for updates what's to stop any closed dictatorship or censorship motivated government from blocking the updates and censoring what's on the box before it's allowed to be used in their territory? Even more alarming is the fact that they could actually make the use of these things illegal.
So are we saying that satellite reception hardware can't be detected remotely? Presumably, some form of handshake takes place along the way so that duplicate data isn't downloaded from the transmission stream to the local cache. I grant you that I'm only speculating, but if a signal is broadcast on a particular frequency then how do you stop that particular frequency from being hijacked? Or there is nobody capable of spoofing a satellite transmission signal to bork your bit of hardware? Even if it's as simple as pretending to tell you to download an OTA firmware update?
Presumably, some form of handshake takes place along the way so ...
My equally ill-informed speculation was the opposite of yours. Instead of 2-way comms, I assumed thay were using what's called a "Digital Fountain". The Wikipedia article is a little bit dry, so in summary the sender periodically sends out a packet of data with a header that tells the receiver "this packet is the XOR sum of blocks a [, b, [...]] of the file"*. The receiver will eventually have enough packets to reconstruct the file. Being able to decode the file is probabilistic, with the probability tending towards 1 the more random packets you receive.
This is different from traditional error-correcting codes in several ways, but the main thing is that the sender picks a random selection of blocks each time it sends a packet and just XORs them together (technically, it's "stateless" because it doesn't need to remember what packets it has already sent). So long as the receiver knows how to decode the packet header, any sufficiently higher number of received packets will be enough to recover the full file. So it doesn't matter if you've got shitty reception or can't keep up with the sender sending stuff too fast; you just have to wait a bit longer until you've got the magic number of packets.
This seems like it should be a natural application for "Digital Fountains" since it means that the satellite (sender) doesn't have to engage in any handshaking at all with receiving sites (just a secure uplink from the satellite owners) and the ground boxes don't need transmitters (at least not pointing skywards, anyway).
(*) in practice, the header just consists of a seed value for a random number generator. So long as both sides are using the same RNG and algorithms, they'll both agree on which file blocks are being XORed in any particular packet.
One more thing: there's a daemon available for *nix systems called "flamethrowerd" that does something similar on multicast networks, although it doesn't actually use fountain codes.
@Peshman
" but if a signal is broadcast on a particular frequency then how do you stop that particular frequency from being hijacked? Or there is nobody capable of spoofing a satellite transmission signal to bork your bit of hardware? "
If that's really a concern, then the best trick would be to use a parabolic receiver, which is traditionally used to isolate a particular signal from a particular point in line-of-sight -- in common parlance, a "satellite dish".
Obviously not all satellite receivers bother with the parabolic dish now (GPS, most satellite phones) but if interference is a genuine concern (and it will be if you're operating on a band that isn't specifically reserved by the ITU) then you'll be wanting that dish....
Presumably, some form of handshake takes place along the way so that duplicate data isn't downloaded from the transmission stream to the local cache.
There is no handshake with OTA updates to your Freeview set top box; the set top box knows only to commence an update of itself when it has received an update that satisfies the checksum requirements.
Obviously, the design of the boxes and service has to be such that it isn't dependent upon a box receiving every update correctly.
Despite the touchy-feely and somewhat patronizing crap about teaching kids, supporting democracy and otherwise spreading Western greatness to "remote corner[s] of the world", I kind of like the idea of an offline web cache that dubs as a satellite receiver. And it seems far more doable than the "cubesat" nonsense they were touting earlier. Let's see how this one plays out.
Maybe they just never heard of it? I hadn't.
Also I found no straightforward way to download anything. The main content repository seems to be a Dropbox share but you cannot download any sizable part of it (say the English subfolder) from the web, you'd need to install the desktop plug-in for that. Wikipedia for Schools includes a "download" link that actually sends you to a page stating (halfway through a lot of unrelated topics) that "[t]he full Wikipedia for Schools 2013 edition will soon be available to download via BitTorrent. Watch this space for updates..." I didn't find any indication that other resources (say, Khan Academy videos) could be downloaded separately, but given that dead end I didn't look much.
Then there are rather misleading instructions on how to install "RACHEL servers" that are actually just web servers with the project's content loaded, but how to get the damned content in order to put it in the server isn't clearly explained anywhere... Compared to this sorry mess Lantern does seem like a rather tight proposition: order it, turn it on, connect to the wi-fi hotspot, start reading, mind the updates every now and again. So not so much reinventing the wheel as getting it right the second time around?
Whilst there are techological similarities to Rachel, the real difference is that Lantern intends to be a ready-to-use off-the-shelf platform and system and usable by normal people.
In some respects equivalent to the difference between Ubuntu and DIY Linux distribution; whilst I can give someone a Ubuntu LiveCD and be confident they can get it working, I cannot reasonably expect a non-expert to build a working distribution from all its component parts.
Why? Those things will take up a hell of a lot of bandwidth that could be put to better use. Maybe one or two linux distributions might be OK, but then you'll have people wanting to cram more on there under the guise of 'X is there, why not Y?'. Soon you'll have dozens of Linux distros but some very old copies of Wikipedia.
As for the Bitocin chain, why? Its useless if its not updated as close to real-time as possible, since it would open the door to quite a lot of fraud (Similar to the bad check type scams of yester-year)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
