[Tech] Is Freenet blocked by some internet provider in France ?

Florent Daigniere florent.daigniere at laposte.net
Fri Sep 26 10:08:47 BST 2008


On Fri, 2008-09-26 at 08:48 +0800, Daniel Cheng wrote:
> 2008/9/26 Matthew Toseland <toad at amphibian.dyndns.org>:
> > On Thursday 25 September 2008 17:16, sich wrote:
> >> Hello,
> >>
> >> I post here to inform the dev about some problem from one french user (he
> > doesn't speak english). He have send a message on fms board fr.freenet.
> >> The original author is : herb at 5FeJUDg2ZdEqo-u4yoYWc1zF4tgPwOWlqcAJVGCoRv8.
> >> I have copy / paste is message at the end (if some people understand
> > french).
> >>
> >>
> >> Apparently he encounter connection problem between 18H and 23H.
> >
> > After 23H it works again?
> >>
> >> When there is no problem Freenet run fine, emule without encryption to.
> >> But when he have difficulty only emule with encryption can connect.
> >>
> >> You can find some connection stats here :
> >>
> > USK at tK4-kA2nUwT39pA3io4POy78FIKObVxwUDKSoe57~MQ,glwvhnN-YQi6miDYdQOkafBJVyR2cNxIN7AHx1L6ev4,AQACAAE/flog/23/content-7892B290.html
> >>
> >> Herb says that he think that Freenet is filtered. Is Internet Provider
> > probably use QoS.
> >
> > Sounds like they block p2p traffic during peak hours to improve performance
> > for everyone else...
> >>
> >> What can we do to avoid this connection problem ?
> >
> > Well, if encrypted emule CAN connect but Freenet can't, there is something
> > *very* strange happening! What does encrypted emule look like from a network
> > level? SSL?
> 
> 
> emule use "protocol obfuscation",
> tradition emule send a signature byte, followed by the data in plain text.
> 
> In "protocol obfuscation" mode, it begin with some random bytes and
> use that byte to start a DH key exchange. I can't see how this
> stronger then the one using in freenet.
> 
> see http://wiki.emule-web.de/index.php/Protocol_obfuscation#Technical_implementation
> and http://mldonkey.sourceforge.net/EDKObfuscation
> 

As far as I know emule uses TCP for data transferts... that's the
difference. What we could do is to implement a TCP transport... or
better: what I like to call "transport plugins" (a generic framework to
make freenet's traffic transit over virtually anything, including
steganographic transports).




More information about the Tech mailing list