|
Re: Make the Internet uncensorable to intermediate nodes
2010-03-27 00:19:37
On a second thought, why don't we change the angle of this subject?
Nowadays, dual ISP are pretty common. I may have a cable from Verizon
broadband and a wireless connection from AT&T at the same time. Now
suppose I want to run a bandwidth-demanding online service,
OnLive<http://www.onlive.com/> for example, let's say it requires a
10mbps connection, my cable and my wireless each has only a 6mbps
connection.
Can I *combine* the speed of these two connections as one so I can use OnLive?
I think it's a legit problem IETF need to address. If there are more
and more ISP and connection availability due to advance of technology,
people will always seek for a way to exploit a combination potential
of full bandwidth.
Now, anti-censorship is only a by product of the protocol. We can
create a virtual ISP in a single ISP connection, but it's *hard* to
surveillance or censor.
I mistakenly replied to Dave alone not to the list, sorry. So here's
what I previously said few days ago:
On Sat, Mar 20, 2010 at 11:27 AM, MtFBwU
<may(_dot_)4thbwu(_at_)gmail(_dot_)com> wrote:
Thanks for all the replies
Such a censorship system would be quite stupid. We would not even
need complicated protocols to workaround it, just using synonyms
or euphemisms would suffice.
Haha, very true indeed. Now let me tell you a real story
China blocked youtube right? First it does DNS tamper, so I setup a local
DNS server, forward 8.8.8.8 using TCP would solve the problem.
Then the Great Firewall (G.F.W.) do URL blocking, it concatenate the HOST
header and the GET strings together then judge if your HTTP query is
unwelcome. I have also discovered a way to bypass it, we can actually use a
double or multiple space after GET like
GET / HTTP/1.1
HOST: www.youtube.com
I can get partial return of HTML. Because sadly, the last block, DPI checkes
this string
<title>YouTube - Broadcast Yourself.</title>
It would RST you in the middle of a transport. So I can only load HTML up to
the title part
So really, fighting against censorship is a two-way game. Yes we can
use synonyms or euphemisms in domestic communications, we can do double
thinking, but the real problem is we can not ask the other end like youtube
or google to change its fingerprints regularly and accordingly to
each censorship. We have to and we can solve the problem in a lower level,
once and for all.
he resistance in Cuba uses USB thumb drives to transport
information. Looking at ways to improve the use of such drives is likely to
produce a more effective counter-censorship scheme.
You see, the G.F.W is in fact, under a lot of pressure. Evidence shows that
China use a massive cluster of Shuguang 4000L super computer farm to do the
censorship job, if we have a FEC-like protocol, along with multiple
BitTorrent downloading sessions simultaneously open on each client, then the
censorship would not work properly. In the past we have encountered GFW
failure from time to time, because it was during the evening where Internet
activity peaks in China.
In this real-time web world, using USB thumb would be too slow for important
information spreading. :P
What "prior art" research have you done? What did you find, and why
wasn't it suitable? What do you see as the already available building
blocks, or concepts to extend?
I am very sorry, I know IETF is a place where people discuss technical
details, but currently I do not have that comprehensive knowledge to go any
futther. As a user, this thread is more like just a suggestion to you guys,
if there will be an important protocol to be designed for the future, please
consider making it to be intermediate node agnostic.
On Sat, Mar 20, 2010 at 10:21 AM, Dave Aronson
<ietf2dave(_at_)davearonson(_dot_)com>
wrote:
On Fri, Mar 19, 2010 at 09:59, MtFBwU
<may(_dot_)4thbwu(_at_)gmail(_dot_)com> wrote:
I am an average Internet user from China. Sorry for my bad English.
Actually, it seems fairly good to me. Anybody who can understand, let
alone come up with, a username like yours, obviously has a pretty good
grasp of it. :-)
In my opinion, theoretically, we *can* make the Internet uncensorable,
In the large, it already essentially is. Find one tiny little
pinhole, through which to leak something to somewhere free, and it
cannot be erased from the net as a whole. (Note that said pinhole
need not be via the net! Leak it on paper in a bottle, and someone
might find it and post it to the net.) Anything from reports of
power-embarassing events, to the old goatse pix, are still available
SOMEwhere.
The TL;DR answer is FEC algorithms.
Hmmm, interesting. I'm not an info-theory wonk, but at first blush,
late on a Friday evening, this sounds plausible, to me. As Stephane
points out, some of it is already popular. It sounds like you want to
combine the diverse routing of BitTorrent (and ToR?), with some
steganography ("debris nobody will notice", possibly in non-user
data), and FEC to account for the possibility of some data being
blocked or altered.
What "prior art" research have you done? What did you find, and why
wasn't it suitable? What do you see as the already available building
blocks, or concepts to extend?
-Dave
--
Dave Aronson - Have Pun, Will Babble | Work: davearonson.com | /\ ASCII
-------------------------------------+ Play: davearonson.net | \/ Ribbon
"Specialization is for insects." | Life: dare2xl.com | /\ Campaign
-Robert A. Heinlein | Wife: nasjleti.net | Email<>Web
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf
|
|